<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Rutvik Deshpande on Medium]]></title>
        <description><![CDATA[Stories by Rutvik Deshpande on Medium]]></description>
        <link>https://medium.com/@rutvik.desh?source=rss-7970904ad169------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*EDqgTSKChaftJdJZIrLHtw.jpeg</url>
            <title>Stories by Rutvik Deshpande on Medium</title>
            <link>https://medium.com/@rutvik.desh?source=rss-7970904ad169------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sat, 21 Nov 2026 23:09:25 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@rutvik.desh/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Future-proof Neighbourhoods: GPT-powered SWOT Analysis for Urban Design & Planning]]></title>
            <link>https://medium.com/digital-blue-foam/future-proof-neighbourhoods-gpt-powered-swot-analysis-for-urban-design-planning-6ff7b361b3d4?source=rss-7970904ad169------2</link>
            <guid isPermaLink="false">https://medium.com/p/6ff7b361b3d4</guid>
            <category><![CDATA[cities]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[swot-analysis]]></category>
            <category><![CDATA[urban-design]]></category>
            <dc:creator><![CDATA[Rutvik Deshpande]]></dc:creator>
            <pubDate>Mon, 12 Feb 2024 19:54:14 GMT</pubDate>
            <atom:updated>2024-02-12T19:54:14.467Z</atom:updated>
            <content:encoded><![CDATA[<h4>RESEARCH &amp; DEVELOPMENT</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*DCHVO2y_dr351Q_BKn_KCQ.png" /></figure><p><em>Hi, it’s Rutvik and Naomi from DBF. We are passionate about creating evidence-based tools to improve the built environment. Having worked on projects like developing Large Language Models (LLM) for predicting </em><a href="https://medium.com/digital-blue-foam/unlocking-the-power-of-gpt-3-how-we-transformed-it-into-a-15-minute-city-expert-f1abb11593c8"><em>15-minute city scenarios</em></a><em> and </em><a href="https://medium.com/digital-blue-foam/real-time-collaboration-for-transit-oriented-developments-83f87d731feb"><em>gamifying urban developments</em></a><em>, we are excited to share our journey of developing Urban Insight Analysis. In this blog post, we harness the power of GPT to generate actionable insights, aiming to contribute to the creation of better cities through innovative technologies.</em></p><h3>How to move beyond KPIs to design recommendations ?</h3><p>When it comes to evaluating the performance of a neighborhood, having comprehensive quantitative evaluation metrics is undoubtedly valuable. DBF has introduced standardized metrics such as fulfillment score, diversity score, proximity score, CO2 score, and density score, enabling data-driven decision-making within the context of the 15-minute city framework [1]. However, interpreting these results often requires an expert level of knowledge, making it challenging for various stakeholders, including architects, tourists, city planners, developers, and residents, to fully grasp and act upon the findings. This is where text-based insights, in the form of natural language, step in to bridge the gap.</p><p>The idea of integrating Language Language Models (LLMs) to create recommendation engines for neighborhood design ideation has been explored previously [2]. The <strong>GPT-3</strong> base model was fine-tuned with a structured dataset consisting of real world examples to generate design proposals in the form of natural language. One of the key challenges faced was that the generated solutions were often semantically similar due to lack of diversity in the dataset. In this project, the LLM by <a href="https://openai.com/">Open AI</a>, <strong>GPT-3 </strong>[3] was used to interpret the <em>15&quot; City Key Performance Indicators</em> (KPIs) and generate smarter insights &amp; design recommendations rather than fine-tuning the model on a dataset.</p><h4>Initial Experiments</h4><p>To understand how we could leverage chat GPT, we embarked on some initial experiments. We first identified potential use cases in which text-based output can be useful.</p><p>We were curious, how we could translate 15&quot; City metrics can be a valuable insight in the form of natural language. The example is for a neighborhood located in <em>Adelaide, Australia</em>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*K0rkuCYQsqufAmG4QeErDA.png" /><figcaption>15” City Metrics for the neighborhood of Mile End South, Adelaide, Australia</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*wTvmfy8yNwQ4yfvNjAS86g.png" /><figcaption>GPT-generated Output according to the 15” City Metrics</figcaption></figure><p>Text-based insights in urban evaluations provide a deeper understanding of a neighborhood’s performance by capturing its historical, cultural, and social contexts. These insights go beyond numbers, revealing connections and trends that quantitative metrics may miss. By incorporating qualitative aspects, they enrich the evaluation process, promoting a more inclusive and people-centric approach to urban planning and development.</p><p>Following are some of the advantages of text-based insights over numerical metrics:<br>1. <em>Contextual Understanding<br>2. Uncovering Hidden Patterns<br>3. Subjective and Qualitative Aspects</em></p><h3><strong>From Insights to Actions</strong></h3><h4>Generating SWOT analysis using GPT</h4><p>We have explored how we can get actionable insights from the data, and not just qualitative analysis. We referred to the <strong><em>SWOT framework</em></strong> and looked into how we can generate this using GPT. For the purpose of generating SWOT analysis, we used results from 3 analyses (i.e. <a href="https://harmless-jar-93f.notion.site/15-Min-City-Analysis-Guide-220d013830e64ecc8849dcb7666a9526#974d836dde2148b9945dd9280b8f5d0c"><em>Fulfillment</em></a>, <a href="https://harmless-jar-93f.notion.site/15-Min-City-Analysis-Guide-220d013830e64ecc8849dcb7666a9526#8065d8137e2a49519978b79057773d4a"><em>Diversity</em></a><em> </em>and <a href="https://harmless-jar-93f.notion.site/15-Min-City-Analysis-Guide-220d013830e64ecc8849dcb7666a9526#3d69991850cc4bba84abb3752f59cfe7"><em>Proximity</em></a>) that we have developed previously as input into the GPT. In essence, what we did was to build a layer of insight generation, on top of our proprietary data.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*sL_YsA0zyB_NZm4PYhu0rg.png" /><figcaption>3 analyses based on 15” City Framework available in DBF: Fulfillment, Diversity and Proximity</figcaption></figure><p><em>Below is a sample result that we obtained for Singapore.</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*FcZ-FMMbIQcQSPIDD5BNpQ.png" /></figure><p>Here, we used GPT to generate the Strengths, Weaknesses and Opportunities. For the Threats section, we plan to give users access to various climate-related threat analysis such as <strong>flood risk, urban heat island and earthquake risk</strong>. This will be specific to their countries, as different countries would have different dataset available, in different formats.</p><h4>UX Design</h4><p>To make the insights more valuable to users, we believe that Linking the map data to the texts shown is important as it gives a spatial quality to the insights.</p><p>We created a use case where this feature could have a storytelling mode. This can be a tool to assist urban planners and designers in narrating their vision for the neighborhood. Below is a sample we did for <em>Brickell, Miami</em>. The idea is for different features in the map to be highlighted, corresponding to whether Strengths, Weaknesses, Opportunities or Threats are selected.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*z4H7FCY0bqxniazOz_blMA.gif" /><figcaption>Spatial Correlation for SWOT Analysis</figcaption></figure><p>By incorporating spatial qualities to the analysis, users can quickly identify locations which need interventions. For example, users can identify the missing amenities, see how far someone would need to travel beyond the 15” isochrone and start to identify where the new amenities should be introduced.</p><h3><strong>How we can Interpret Urban SWOT Analysis to Improve Neighborhoods?</strong></h3><h4>What to build?</h4><p>When evaluating a neighborhood using metrics like <em>fulfillment score</em> and <em>diversity score</em>, it becomes apparent which essential facilities are lacking. These metrics provide users with a clear understanding of the existing gaps and areas for improvement within the neighborhood. However, identifying these missing facilities is only the first step.</p><p>For example a neighborhood can score low on learning and caring because there aren’t enough schools, libraries, hospitals or clinics to serve the population. Urban insights highlights the top most critical facilities missing.</p><p>The next crucial question is:</p><blockquote><strong>How can we determine the ideal locations for these new facilities?</strong></blockquote><h4>Where to build?</h4><p>To address this challenge, we propose a rule-based algorithm that facilitates the identification of optimal locations for the new facilities. The algorithm incorporates default rules centered around factors such as <strong>proximity to public transportation (<em>including metro stations and bus stops</em>), existing similar facilities, street connectivity and risks (e.g. flood risk and urban heat island). </strong>Then the algorithm rates suitable locations based on these rules and presents the most ideal places to build this facility.</p><h4>Measuring Impact</h4><h3><strong>Conclusion</strong></h3><p>We explored how to leverage GPT and draw insights from how many? data points. There is value in being able to get SWOT analysis quickly for different sites. It becomes more interesting when we start to make spatial correlations to the texts. Meanwhile, location recommendations can reduce the time required for users to select a site and could potentially choose better sites than manual selection too. All this will contribute to the overall mission of designing better cities.</p><h3><strong>Next Steps</strong></h3><p>So far we have used the DBF analysis data to act as input in the generation of insights. In the future we can look into augmenting it with other data sources such as demographic data, land use and even qualitative data from Wikipedia. The foreseeable challenge here is the inconsistency of data (formats and availability) across geographical locations. One way to deal with this is to narrow the scope of the exploration to a specific city or country, and slowly expand to cover more geographical areas.</p><p>While leveraging the power of <strong><em>GPT-3.5</em></strong> has been instrumental in generating insightful analyses from DBF data, it’s essential to acknowledge certain technical limitations. It may not have access to the most up-to-date information or specific localized data, which can present challenges when augmenting the analysis with additional data sources.</p><h3><strong>NOTES</strong></h3><h4>SWOT Analysis</h4><p>The term SWOT refers to Strengths, Weaknesses, Opportunities and Threats. This approach allows urban designers and planners to assess the existing conditions and identify potential interventions and strategies to enhance the neighborhood’s livability and functionality [4]. SWOT analysis provides a valuable framework for urban analysis, facilitating a systematic and structured approach to understanding the intricacies of a site and informing the subsequent design process.</p><h4>DBF Urban Insights</h4><p>Meanwhile, the <a href="https://www.c40knowledgehub.org/s/article/Carlos-Moreno-The-15-minute-city?language=en_US">15” City Framework</a> proposes an urban planning methodology where inhabitants have access to all the amenities they need within a walking distance [5]. DBF has developed its <a href="https://www.digitalbluefoam.com/feature/urban-insights">urban insights capabilities</a> which allows users to get objective evaluation on how well the neighborhood is performing as a 15” City neighborhood. These capabilities were our first attempt in developing an objective evaluation method of a neighborhood. We wanted to take this feature a step further, and look at how we can translate these numerical data into actionable insights.</p><h4><strong>CITATIONS</strong></h4><ol><li>Cheng, C., Hsain, H. E., Deshpande, R., Nisztuk, M., Chavan, T., Patel, S., &amp; Weijenberg, C. (2022). DBF Urban Insights: Realtime 15-Minute City Neighbourhood Analysis System.</li><li>Deshpande, R., Vijay Patel, S., Weijenberg, C., Nisztuk, M., Corcuera, M., Luo, J., &amp; Zhu, Q. (2023). Generative Pre-Trained Transformers for 15-Minute City Design. 595–604. <a href="https://doi.org/10.52842/conf.caadria.2023.1.595">https://doi.org/10.52842/conf.caadria.2023.1.595</a></li><li>Brown, T., Mann, B., Ryder, N., Subbiah, M., Kaplan, J. D., Dhariwal, P., Neelakantan, A., Shyam, P., Sastry, G., Askell, A., Agarwal, S., Herbert-Voss, A., Krueger, G., Henighan, T., Child, R., Ramesh, A., Ziegler, D., Wu, J., Winter, C., … Amodei, D. (2020). Language Models are Few-Shot Learners. Advances in Neural Information Processing Systems, 33, 1877–1901. <a href="https://papers.nips.cc/paper/2020/hash/1457c0d6bfcb4967418bfb8ac142f64a-Abstract.html">https://papers.nips.cc/paper/2020/hash/1457c0d6bfcb4967418bfb8ac142f64a-Abstract.html</a></li><li>Grisiute, A., Shi, Z., Chadzynski, A., Silvennoinen, H., von Richthofen, A., &amp; Herthogs, P. (2022). Automated Semantic SWOT Analysis for City Planning Targets: Data-driven Solar Energy Potential Evaluations for Building Plots in Singapore. 555–564. <a href="https://doi.org/10.52842/conf.caadria.2022.1.555">https://doi.org/10.52842/conf.caadria.2022.1.555</a></li><li>Moreno, C., Allam, Z., Chabaud, D., Gall, C., &amp; Pratlong, F. (2021). Introducing the “15-Minute City”: Sustainability, Resilience and Place Identity in Future Post-Pandemic Cities. Smart Cities, 4(1), Article 1. <a href="https://doi.org/10.3390/smartcities4010006">https://doi.org/10.3390/smartcities4010006</a></li></ol><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=6ff7b361b3d4" width="1" height="1" alt=""><hr><p><a href="https://medium.com/digital-blue-foam/future-proof-neighbourhoods-gpt-powered-swot-analysis-for-urban-design-planning-6ff7b361b3d4">Future-proof Neighbourhoods: GPT-powered SWOT Analysis for Urban Design &amp; Planning</a> was originally published in <a href="https://medium.com/digital-blue-foam">Digital Blue Foam</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Unlocking the Power of GPT-3: How We Transformed it into a 15-Minute City Expert]]></title>
            <link>https://medium.com/digital-blue-foam/unlocking-the-power-of-gpt-3-how-we-transformed-it-into-a-15-minute-city-expert-f1abb11593c8?source=rss-7970904ad169------2</link>
            <guid isPermaLink="false">https://medium.com/p/f1abb11593c8</guid>
            <category><![CDATA[urban-design]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[gpt-3]]></category>
            <category><![CDATA[nlp]]></category>
            <dc:creator><![CDATA[Rutvik Deshpande]]></dc:creator>
            <pubDate>Thu, 19 Jan 2023 04:33:13 GMT</pubDate>
            <atom:updated>2023-01-19T04:33:13.086Z</atom:updated>
            <content:encoded><![CDATA[<h4>RESEARCH &amp; DEVELOPMENT</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*lzZhZMqxuQxSszHdQ9ofmw.png" /><figcaption>Can we transform GPT-3 into a 15-minute city expert?</figcaption></figure><h3><strong>GPT-3 Meets Urban Design</strong></h3><p>With the release of <em>ChatGPT </em>[1]<em> </em>and <em>Midjourney,</em> 2022 was a year where humanity made big leaps towards <em>Artificial General Intelligence (AGI) </em>will shape our near future. Like many companies and researchers, we at DBF are curious about the potential of these tools to shape new ways of working in our industry.</p><p>In this blog, we discuss how we are exploring the potential of an AGI called <em>GPT-3</em> through the lens of a building and city design. The <em>Generative Pre-trained Transformer 3 (GPT-3)</em> is an autoregressive language model that leverages deep learning to produce human-like text [2]. Specifically, we look at how we can use a method called <em>transfer learning</em> and <em>GPT-3</em> to build an AI which is an expert on a recent urban design trend called the <em>“15-Minute City”</em>, and how this is helping us to rethink our company mission to facilitating the design of more planet and human friendly buildings and cities.</p><h3><strong>What is a 15-Minute City?</strong></h3><p>The <em>15-Minute City</em> is a residential urban concept in which most basic amenities are available by either walking or cycling from residents’ homes, resulting in pedestrian-friendly neighborhoods, reducing CO2 emissions, and improving the quality of life.[3]</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*lkRFrq55rR5pCBfx2cVWmw.png" /><figcaption><em>The 15-Minute City illustration; </em>Source: <a href="https://www.weforum.org/agenda/2022/03/15-minute-city-stickiness/">The surprising stickiness of the “15-minute city”</a></figcaption></figure><p>Cities globally are adopting The 15-Minute City model as an urban response to the crises. Though the ability to explore and validate <em>15-Minute City</em> designs is difficult for several reasons:</p><ul><li>Urban design must consider location-specific factors such as climate, demographics, and culture for optimal results.</li><li>A holistic approach involving multiple experts from various disciplines is key to creating sustainable and livable urban areas.</li><li>The use of standardized evaluation criteria, such as the 15-Minute City principle, ensures the design meets the needs of the residents efficiently and effectively.</li></ul><h3><strong>How can GPT-3 help us design 15-Minute Neighborhoods ?</strong></h3><p>In this project, we explore the use of <em>GPT-3</em> to support designers to come up with novel and feasible concepts to link location-specific requirements with potential design solutions in urban design by introducing a novel “<em>design concept generation</em>” tool that utilizes <em>Pre-trained Language Models (PLMs)</em>. Our research aims to explore the potential of PLM, specifically <em>Large Language Models (LLMs)</em>, to generate recommendation engines for neighborhood design ideation in the form of natural language. This tool is specifically designed for the 15 Minute City concept, which aims to create urban areas where residents have access to essential goods and services within a 15-minute walk or bike ride.</p><p>We used a data-driven approach to retrieve data from the C40 CITIES database, fine-tune an LLM, <em>GPT-3</em>, and generate location-specific neighborhood design solutions based on given conditions. Making this AI system available to designers through a recommendation system would augment designers’ creativity and enhance their knowledge base.</p><h3><strong>Making it happen</strong></h3><p>The method to develop a virtual expert based on 15-Minute City using <em>PLM </em>consists of four parts:</p><ol><li>Collecting a structured set of urban design examples</li><li>Pre-processing the data and creating input-output pairs of text</li><li>Fine-tuning the base PLM with a new set of examples</li><li>A web-based user interface that allows interaction with this virtual expert</li></ol><h4><strong>Data Hunting</strong></h4><p>Data collection was the most important step in this project. The textual dataset we used to fine-tune the base PLM was collected from winning entries of <em>“Reinventing Cities”</em> competition conducted by <a href="https://www.c40.org/">C40 Cities</a> [4]. The competition aimed to accelerate the development of low-carbon, resilient urban regeneration. The proposals were evaluated by city mayors, developers, and urban design experts based on the project’s ability to serve as a model for carbon-neutral development, providing local community benefits, and innovative and replicable climate solutions.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*VaXdx0yZ5MxCHSwfauXGfw.png" /><figcaption><em>The data was collected from the “Reinventing Cities” competition from C40 CITIES database</em></figcaption></figure><p>The competition brief for each city includes textual information like site description, expected program, plot area, type of property transfer intended, site context and redevelopment goals, climate risks, environmental challenges, and planning rules and regulations, for neighborhoods from cities around the world. The proposed solutions of the winners included a solution description, key components and solutions, and a presentation of the site.</p><h4><strong>Technical Execution</strong></h4><p>However, it may become challenging to fine-tune these models for specific tasks. Few-shot learning is a machine learning technique that enables a pre-trained model to generalize over new categories of data using only a few labeled samples. Because PLMs like <em>GPT-3</em> are capable of few-short learning, they only needed a few innovative, rational, and validated examples of neighborhood design proposals.</p><p>This competition had 33 excellent winning solutions based on site requirements from around the world. The textual data was reformulated and customized to create different input-output pairs. The input text consisted of the combination of site location, climatic conditions, expected program, and the total site area. The output text was the following set of topics, submitted by the winners :</p><ol><li>Key Components</li><li>Solution Description</li><li>Presentation of the site</li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*D9mh0hfMRwckuKk9BPLMmw.png" /><figcaption>The input prompt structure used for fine tuning the base model, consisting of site location, climate description, expected program, and site area</figcaption></figure><p>The base <em>GPT-3</em> model was fine-tuned for different tasks to better understand the context and constraints. The structured dataset was used for fine-tuning the base model, with the API offered by <a href="https://openai.com/research/">OpenAI</a> [5].</p><h4>User Experience</h4><p>Looking at the massive success of <em>ChatGPT</em>, it is clear that providing a natural and intuitive experience for users is crucial to make an impact. We created a simple web based user interface to facilitate this approach of generating neighborhood design solutions. The user can input different parameters as text and click on generate to get novel and feasible neighborhood design solutions.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*S8eCEbi-HUZsL5Y8wZb7lw.gif" /><figcaption><em>A simple and intuitive user interface of web based 15 Minute City AI tool, to generate neighborhood design concepts based on the user inputs</em></figcaption></figure><h3><strong>Results</strong></h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*1B8osnGLEob_fP1Nb6Ngfw.png" /><figcaption>The fine-tuned model generating “key components” for a solution based on given input text</figcaption></figure><h3><strong>Evaluation Criteria</strong></h3><p>These generated design solutions were assessed based on an evaluation metric inspired by the work of Data-Driven Innovation Lab at the Singapore University of Technology and Design. They developed this rating methodology to evaluate bio-inspired design ideas using pre-trained language models [6]. A survey was conducted with designers and architects involved to evaluate the generated neighborhood design solutions based on the rating rubrics, consisting of <em>feasibility </em>and <em>novelty </em>on the solutions. We compared feasibility and novelty scores of the solutions generated by different base <em>GPT-3</em> models provided by <a href="https://beta.openai.com/docs/models/gpt-3">OpenAI</a>; <em>Ada, Babbage, and DaVinci</em>. The rubrics used for the evaluation for generated neighborhood design solutions is shown in tables below.</p><pre>┌─────────────┬────────────────────────────────────────────────────────────────────────────────────────────────────────┐<br>│ Feasibility │ Description                                                                                            │<br>│ Rank        │                                                                                                        │<br>├─────────────┼────────────────────────────────────────────────────────────────────────────────────────────────────────┤<br>│ 1           │ The solutions make no sense from the 15 minute city perspective                                        │<br>│ 2           │ The solution makes little sense with today’s technology, but could be possible in the future           │<br>│ 3           │ The solution makes sense, but efforts are needed to work out a design roadmap                          │<br>│ 4           │ The solution makes good sense, and a design roadmap can be easily established to realize it            │<br>│ 5           │ The solution makes perfect sense and there are existing resources, materials, or methods to realize it │<br>└─────────────┴────────────────────────────────────────────────────────────────────────────────────────────────────────┘</pre><pre>┌─────────────┬────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────┐<br>│ Novelty     │ Description                                                                                                                                        │<br>│ Rank        │                                                                                                                                                    │<br>├─────────────┼────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────┤<br>│ 1           │ Neighborhood solution exists and is commonly seen in the cities                                                                                    │<br>│ 2           │ Neighborhood solution exists but is uncommon in terms of a 15-Minute City                                                                          │<br>│ 3           │ New components and metrics are proposed for a 15-Minute City, but similar approaches or technology can be commonly seen in related urban contexts  │<br>│ 4           │ New features are proposed for a 15-Minute City, and similar approaches or technology can be rarely found in related urban contexts                 │<br>│ 5           │ New features are proposed, and no similar design strategies, approaches or technology can be found nowadays                                        │<br>└─────────────┴────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────┘</pre><p>The results generated by <em>Ada</em> and <em>Babbage</em> base models got a good feasibility score but lower novelty score, while the ones generated from the <em>Davinci</em> model were considered more feasible as well as novel. Based on the generated solutions, the model seems to understand the intent of the text.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/720/1*aVWnwC6KAmgoQvPCWINY_w.png" /><figcaption>Human evaluation results based on the predefined rubrics for feasibility scores and novelty scores</figcaption></figure><h3><strong>Next Steps</strong></h3><p>One of the main challenges in this approach was finding a well-structured dataset following the <em>15 Minute City</em> principles. While this project explored the generative aspect of urban designing using the <em>GPT-3</em>, the next step would be creating an evaluation engine based on the <em>15-Minute City</em> framework and logic developed by urban design experts, to evaluate and rank the generated design solutions.</p><h3><strong>Final thoughts</strong></h3><p>In this project, we tried to explore the potential of PLMs to create recommendation engines for neighborhood design ideation processes in the form of natural language. These language models have powerful natural language understanding capabilities, the model not only interprets and understands the intent of the text but can also generate novel and feasible design ideas for a <em>15-Minute City</em> in the form of natural language that is semantically similar to that of humans.</p><p>We want to take the impact of <em>GPT-3</em> goes far beyond chat interfaces. With its capacity to generate and analyze vast amounts of data, it is opening up new possibilities for designers to create more efficient and effective user experiences. The future of design is no longer limited by our human imagination, the LLMs are now enabling us to dream up designs that were once thought to be impossible.</p><h3>References</h3><p>[1] OpenAI, “ChatGPT: Optimizing Language Models for Dialogue,” <a href="https://openai.com/blog/chatgpt/">https://openai.com/blog/chatgpt/</a></p><p>[2] T. Brown et al., “Language Models are Few-Shot Learners,” in Advances in Neural Information Processing Systems, 2020, vol. 33, pp. 1877–1901. Accessed: Oct. 31, 2022. [Online]. Available: <a href="https://papers.nips.cc/paper/2020/hash/1457c0d6bfcb4967418bfb8ac142f64a-Abstract.html">https://papers.nips.cc/paper/2020/hash/1457c0d6bfcb4967418bfb8ac142f64a-Abstract.html</a></p><p>[3] C. Moreno, Z. Allam, D. Chabaud, C. Gall, and F. Pratlong, “Introducing the ‘15-Minute City’: Sustainability, Resilience and Place Identity in Future Post-Pandemic Cities,” Smart Cities, vol. 4, no. 1, Art. no. 1, Mar. 2021, doi: 10.3390/smartcities4010006.</p><p>[4] C40 | Reinventing Cities, <a href="https://www.c40reinventingcities.org/en/">https://www.c40reinventingcities.org/en/</a></p><p>[5] OpenAI API Documentation, “Fine-tuning,” <a href="https://beta.openai.com/docs/guides/fine-tuning">https://beta.openai.com/docs/guides/fine-tuning</a>.</p><p>[6] Q. Zhu, X. Zhang, and J. Luo, “Generative Pre-Trained Transformers for Biologically Inspired Design,” presented at the ASME 2022 International Design Engineering Technical Conferences and Computers and Information in Engineering Conference, Nov. 2022. doi: <a href="http://10.0.4.91/DETC2022%E2%80%9390366">10.1115/DETC2022–90366</a>.</p><h4>About DBF</h4><p><a href="https://www.digitalbluefoam.com/">Digital Blue Foam</a> (DBF) comprises an elite mix of designers and technologists from around the world who share a strong commitment to empowering a revolution in architecture, engineering, and construction (AEC) industries toward carbon-negative projects by leveraging data-driven, AI-powered, collaborative, and sustainable approaches. We embrace collaboration and sponsorship, and we thrive at offering customized solutions that make designing a hassle-free and intuitive process.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=f1abb11593c8" width="1" height="1" alt=""><hr><p><a href="https://medium.com/digital-blue-foam/unlocking-the-power-of-gpt-3-how-we-transformed-it-into-a-15-minute-city-expert-f1abb11593c8">Unlocking the Power of GPT-3: How We Transformed it into a 15-Minute City Expert</a> was originally published in <a href="https://medium.com/digital-blue-foam">Digital Blue Foam</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Designing the 15 Minute City with AirBnB Reviews and Semantic AI]]></title>
            <link>https://medium.com/digital-blue-foam/designing-the-15-minute-city-with-airbnb-reviews-and-semantic-ai-c6d1c490fc32?source=rss-7970904ad169------2</link>
            <guid isPermaLink="false">https://medium.com/p/c6d1c490fc32</guid>
            <category><![CDATA[cities]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[design]]></category>
            <category><![CDATA[nlp]]></category>
            <dc:creator><![CDATA[Rutvik Deshpande]]></dc:creator>
            <pubDate>Mon, 07 Mar 2022 03:50:15 GMT</pubDate>
            <atom:updated>2022-03-07T03:52:46.976Z</atom:updated>
            <content:encoded><![CDATA[<h4>ML APPLICATIONS</h4><h4>How Semantic AI helps city designers move beyond the objective data to draw insights from experiential data sources.</h4><p><strong>By Rutvik Deshpande + Sayjel Vijay Patel</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Jo8DZAdmtTsk4uuyYgGXsA.png" /><figcaption><strong>Image by Author</strong></figcaption></figure><p>The creation of all-knowing <em>artificial general intelligence</em> (AGI) [1], like <em>J.A.R.V.I.S</em> from Iron Man, remains years away. However, great progress is being made in one aspect of AGI focusing on consuming, interpreting, and classifying information shared by humans called <em>Semantic AI </em>[2].</p><h3>Semantic AI</h3><p>Semantic AI unites two branches of computer science. The first is <em>Natural Language Processing (NLP) — </em>an area of computer science looking at modeling how humans understand spoken words and text. The second is <em>knowledge-graphs</em> (or semantic networks); a particular way of representing and storing information as interlinked entities.</p><p>At <em>Digital Blue Foam</em> (DBF), we are pioneering a new application of semantic AI to help our users inform the design of the <em>15-minute city</em> — a rising trend in urban planning, towards arranging all necessary amenities and services within a 15-minute walking distance for all residences.</p><h3>Towards Semantic Neighborhood Design</h3><p>Recently, we introduced a <a href="https://www.digitalbluefoam.com/whitepapers/neighborhood-scoring-system">web-based tool</a> to help designers, planners, and developers to index how well a specific location performs as a 15-minute city with a <em>Neighborhood Quality Score (NQS)</em>. For this project, the biggest challenge was to hunt down, collect, and verify location-specific data to feed the analysis.</p><p>In the first version of the application, we made use of existing map APIs such as <em>OpenStreetMaps </em>and <em>Google</em>. However, to provide a truly comprehensive solution, relying on these <em>pre-cooked </em>data sources presents some profound issues:</p><ul><li>Data-sets are often incomplete or not up to date</li><li>Data-sets focus on objective criteria such as address, building shape w</li><li>The information does not tell us what is important about a location (ie. is it <em>touristy, quiet, historic</em>)</li><li>Global Information APIs can be cost-prohibited</li></ul><h3>Case Studies: AirBnB and the 15 Minute City</h3><p>Our team recognized quickly a need for our contextual data sources to improve the accuracy and potency of our Neighborhood Quality Index. While researching options, we came across a radical solution — you guessed it, Semantic AI! We determined Semantic AI would allow us to extract neighborhood-specific keywords and summaries from alternative datasets such as social media reviews. This could help us evaluate neighborhood quality by helping us instantly collect qualitative data for a location.</p><h4>Step 1; Hunting and Gathering Data</h4><p>After defining the problem statement, the next step was to hunt and gather relevant sources. Typically this is a tedious, costly, and time-consuming process for urban designers. We explored several different APIs, such as foursquare, and Flickr, and ultimately chose AirBnB’s open-source dataset <a href="http://insideairbnb.com/get-the-data.html">Inside Airbnb</a>. This is because the information AirBnB collects focuses on practical, location-specific, insights to help their customers rate, rank, and compare different properties based on criteria such as proximity to public transport stations, noise levels, parks, , and other important characteristics. In turn, user reviews can create a better perception of neighborhoods, districts, and even cities. An additional benefit is the global availability of these datasets, which are crowd-sourced from individual reviews and opinions.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*N9Leo8N1S6YeXKHZaiFDOQ.png" /><figcaption><strong>User Interface (UI) of Airbnb’s platform, showing the details of listings, which could help users summarize a neighborhood. Image Source : </strong><a href="https://www.airbnb.co.in/"><strong>Airbnb</strong></a></figcaption></figure><p>The dataset includes detailed listings and reviews in a tabular format and the neighborhoods of Singapore in GeoJSON format. Since our project focuses on mining insights from qualitative inputs, we extracted features like <strong>“<em>Neighborhood Overview ‘’ </em></strong><em>and</em> <strong>“<em>Guest Review</em>”</strong> for further analysis. These text-based features could be even correlated with other numerical features like the number of reviews, price, and different review scores, to get an in-depth analysis of a particular location.</p><h4>Step 2; Summarizing User Reviews</h4><p>Text summarization is one of the use cases of Natural Language Processing, referring to highlighting and framing important information from a large chunk of text or document. The next step: Airbnb’s user-generated reviews combined with neighborhood overviews written by guests, travelers, and hosts, for different neighborhoods of Singapore, were summarized in a shorter paragraph of 5–7 lines. This task was done using <a href="https://www.nltk.org/">Natural Language Toolkit (NLTK)</a>(NLTK), an open-source collection of python libraries, programs, and resources for building NLP models.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*2GJJzJdboiJScsApbxJbiw.png" /><figcaption><strong>Model Pipeline showing the process of converting merged text data into meaningful summarized paragraph. Image by Author</strong></figcaption></figure><h4>Step 3; Summarizing and Visualizing Reviews</h4><p>In the last step, we concatenate the reviews for different neighborhoods, and then remove special characters, extra spaces, and stop-words. Next, we <strong><em>tokenize</em></strong> the text; this means converting the words into numbers while considering the position of words in sentences. We then measured the occurrence and frequency of keywords and passages within the reviews. Finally, we extracted the most frequent sentences and keywords to create a summary of the most common elements of a location. The resulting summary informs a more qualitative understanding of the neighborhood in contrast to more objective map data.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/900/1*ezLnSrRIlWP3PkYKFMEKVA.png" /><figcaption><strong>Kallang, the centrally located popular neighborhood, has the most Airbnb listings and reviews in Singapore. Image by Author</strong></figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/960/1*aNm_p9qxdLMyoQ9NqKx9fA.png" /><figcaption><strong>The summarized information about the Novena neighborhood, Singapore, displayed on the Map page of the web-based DBF tool. Image by Author</strong></figcaption></figure><h3>How this Helps City Designers</h3><p>Datasets like OpenStreetMaps rely on crowd-sourcing of information; often, for some cities, information could be sparse or out of date. Furthermore, paid datasets like Global Information System (GIS) focus on numerical or geometric attributes, such as buildings and streets, but fail to recognize hotspots and popular places — something easily retrieved from social media, real estate listings, city guides, and reviews data. Furthermore, Semantic AI unlocks new possibilities to study qualitative factors for a specific location such as walkability, human perception, urban diversity, safety, serenity, cleanliness, and sense of place.</p><p>In this new data layer, user experience is translated to feedback and reviews for many other users, adding another dimension to urban analytics. We believe, these insights will help city designers choose the right building type and function to enhance a particular location.</p><blockquote><em>We believe these insights will help city designers choose the right building type and functions to enhance a particular location.</em></blockquote><h4>What’s Next?</h4><p>This approach is not only limited to Airbnb reviews, and can be expanded by combining more relevant datasets from different domains as mentioned above in Figure 5. The given sources can add in more information:</p><ul><li><strong>Information Services</strong>, ex, Wikipedia: Could offer information on a broader level like transit stations nearby, history of the neighborhood, demographic information.</li><li><strong>Property &amp; Real-estate Platforms </strong>like Zillow, PropertyGuru, Mogul: would be helpful in tracking property value trends, along with assessing the quality of listings in a particular neighborhood.</li><li><strong>Places Guides:</strong> Yelp, Tripadvisor, Foursquare, Google Places, and maps: This can be used to retrieve the popularity of amenities, such as ratings and reviews of restaurants, theaters.</li><li><strong>Transit &amp; Movement:</strong> Grab, Uber: Uber released a public <a href="https://movement.uber.com/?lang=hi-IN">dataset</a> consisting of billions of trips, and calculated travel times, and prices. This could help users understand the real-time taken to reach different neighborhoods from a given location for 100s of cities.</li><li><strong>Social Media reviews</strong>, Reddit, Twitter: Tweets and posts could eventually help us understand the hotspots in each locality, understanding the correlation of emotions and public spaces.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/960/1*-EVFeVWCP8tuO1YWNBmqoQ.png" /><figcaption><strong>Geolocated text data (user reviews) can also be retrieved from various other open-source resources and platforms. Image by Author</strong></figcaption></figure><h3>Conclusion</h3><p>In this post, we focused on one example, specifically using Airbnb reviews to inform 15-minute city design, however, this can be expanded to new applications within urban design and planning; for example, Google Street View Images (SVI) can be used to visually inspect and compare the composition of different architectural and natural elements, like streets, trees and cars, within a particular neighborhood. Growing accessibility to Semantic AI helps new people, like urban designers, get useful insights from previously impenetrable data-sets, like social media reviews. These insights can help inform better decisions, especially at the neighborhood scale.</p><p><strong>N O T E S</strong></p><p><strong>1</strong> <em>An Executive Primer on Artificial General Intelligence</em> | McKinsey. <a href="https://www.mckinsey.com/business-functions/operations/our-insights/an-executive-primer-on-artificial-general-intelligence.">https://www.mckinsey.com/business-functions/operations/our-insights/an-executive-primer-on-artificial-general-intelligence.</a></p><p><strong>2</strong> <em>Why Machine Learning Needs Semantics, Not Just Statistics.</em> Forbes, <a href="https://www.forbes.com/sites/kalevleetaru/2019/01/15/why-machine-learning-needs-semantics-not-just-statistics/">https://www.forbes.com/sites/kalevleetaru/2019/01/15/why-machine-learning-needs-semantics-not-just-statistics/.</a></p><p><strong>3</strong> <em>Six Core Aspects of Semantic AI — DataScienceCentral.Com.</em> Data Science Central, 14 May 2018, <a href="https://www.datasciencecentral.com/six-core-aspects-of-semantic-ai/">https://www.datasciencecentral.com/six-core-aspects-of-semantic-ai/.</a></p><p><strong>S O U R C E S</strong></p><p>Getting started with Google SVI API; <a href="https://developers.google.com/maps/documentation/javascript/streetview">https://developers.google.com/maps/documentation/javascript/streetview</a></p><p>Inside Airbnb. Adding Data to the Debate; <a href="http://insideairbnb.com/">http://insideairbnb.com/</a></p><p>Foursquare Check-ins Dataset; <a href="https://sites.google.com/site/yangdingqi/home/foursquare-dataset">https://sites.google.com/site/yangdingqi/home/foursquare-dataset</a></p><p>Wikipedia Geosearch; <a href="https://www.mediawiki.org/wiki/API:Geosearch">https://www.mediawiki.org/wiki/API:Geosearch</a></p><p>Uber Movement Data; <a href="https://movement.uber.com/?lang=hi-IN">https://movement.uber.com/?lang=hi-IN</a></p><p>Zillow Housing Data; <a href="https://www.zillow.com/research/data/">https://www.zillow.com/research/data/</a></p><p><strong><em>About the Authors</em></strong></p><p><strong>Rutvik Deshpande</strong> is a ML Design Engineer at Digital Blue Foam. He aims to blend Artificial Intelligence in Architecture, with his work focusing on data-driven design workflows in both urban and architectural scales. Rutvik has co-led technical workshops at prestigious global conferences like, CAADRIA and DigitalFutures.</p><p><strong>Sayjel Vijay Patel</strong> is the co-founder and CTO of Digital Blue Foam and founding Assistant Professor at the Dubai Institute of Design and Innovation. He is a graduate of the MIT School of Architecture and Planning. Sayjel won the acclaimed Red-Dot Design Award for his research developing conceptual design software for the 3D printing industry.</p><p><em>About Digital Blue Foam</em></p><p><a href="https://www.digitalbluefoam.com/">Digital Blue Foam</a> (DBF) comprises an elite mix of designers and technologists from around the world who share a strong commitment for empowering a revolution in architecture, engineering, and construction (AEC) industries toward carbon-negative projects by leveraging data-driven, AI-powered, collaborative, and sustainable approaches. We embrace collaboration and sponsorship, and we thrive at offering customized solutions that make designing a hassle-free and intuitive process. To learn more about Digital Blue Foam, visit our <a href="http://www.digitalbluefoam.com/">website</a>.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c6d1c490fc32" width="1" height="1" alt=""><hr><p><a href="https://medium.com/digital-blue-foam/designing-the-15-minute-city-with-airbnb-reviews-and-semantic-ai-c6d1c490fc32">Designing the 15 Minute City with AirBnB Reviews and Semantic AI</a> was originally published in <a href="https://medium.com/digital-blue-foam">Digital Blue Foam</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Using Artificial Intelligence to Predict Building Daylight Autonomy]]></title>
            <link>https://medium.com/digital-blue-foam/using-artificial-intelligence-to-predict-building-daylight-autonomy-412166369a3d?source=rss-7970904ad169------2</link>
            <guid isPermaLink="false">https://medium.com/p/412166369a3d</guid>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[sustainability]]></category>
            <category><![CDATA[sustainable-development]]></category>
            <category><![CDATA[architecture]]></category>
            <dc:creator><![CDATA[Rutvik Deshpande]]></dc:creator>
            <pubDate>Wed, 16 Dec 2020 17:46:36 GMT</pubDate>
            <atom:updated>2020-12-17T17:20:48.626Z</atom:updated>
            <content:encoded><![CDATA[<h4><strong>Sustainable Digitization</strong></h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*J1PWMLB4pNTPuVZ8EiqQeg.jpeg" /></figure><p>The goal of sustainable design is to reduce the negative environmental impacts of buildings while boosting the health and comfort of building occupants. While architects design buildings to be comfortable, too often this is dependent on energy-intensive systems such as heating, air-conditioning, ventilation, and lighting.</p><p>In this article, we discuss recent research looking at using a data-driven approach that uses Machine Learning to predict daylighting for building layouts. Compared to other simulation-based methods, this approach allows us to instantly analyze thousands of design options in seconds to determine the best fenestration and building layout to maximize daylighting for a given project.</p><h3>Daylighting Principles</h3><p>One of the key factors in a <em>sustainable design</em> is <strong>“Daylighting”. </strong>Daylight plays a critical role in sustainable building strategies because it is a highly cost-effective and renewable strategy to reduce energy use, cost, and boost occupant well-being, visual comfort, productivity, and health.</p><p>Some of the key strategies that are followed by designers and architects to daylight into design efficiently are: optimizing building orientations for maximum sun exposure, using correct indoor and outdoor flooring materials to achieve the right reflectance, placement of windows, light shafts, skylights, and atriums.</p><p>Here are the rules of thumb for daylighting in a design<a href="https://ocw.mit.edu/courses/architecture/4-401-environmental-technologies-in-buildings-fall-2018/lecture-slides-1/MIT4_401F18_lec8.pdf">[1]</a>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*X1V5YTM8GATL-Amp0Nm2ZQ.png" /><figcaption>Image source : <a href="https://ocw.mit.edu/courses/architecture/4-401-environmental-technologies-in-buildings-fall-2018/lecture-slides-1/MIT4_401F18_lec8.pdf">Daylighting Design Principles</a></figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*An_mAKNdGkDmgKbe1vKW4g.png" /><figcaption>In order to daylight all spaces bordering interior of atrium with diffused daylight, maximum atrium height is about 2.5 it width ; Image source : <a href="https://ocw.mit.edu/courses/architecture/4-401-environmental-technologies-in-buildings-fall-2018/lecture-slides-1/MIT4_401F18_lec8.pdf">Daylighting Design Principles</a></figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*fHckCygb4gkKarGO3KmP4w.png" /><figcaption>In a side-lit space with a standard window and venetian blinds, the depth of the daylit area usually lies between 1.5 and 2 times the window-head height ; Image source : <a href="https://ocw.mit.edu/courses/architecture/4-401-environmental-technologies-in-buildings-fall-2018/lecture-slides-1/MIT4_401F18_lec8.pdf">Daylighting Design Principles</a></figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*srDsqLrkM2SL1HXY6TKOXw.png" /><figcaption>While considering tight urban context, minimum 3 meters should be the offset from the street and there should be a raised overhang of min. 9 meters; Image source : <a href="https://www.digitalbluefoam.com/">Digital Blue Foam</a></figcaption></figure><p>In this experiment, the idea was to develop a data-driven process to design spaces with sufficient daylight.</p><h3>The Data Approach</h3><p>The daylighting data was generated using <em>Ladybug </em>and <em>Honeybee</em>, the environmental plug-ins for <em>Grasshopper3D</em>.<strong> </strong>Parameters provided for the Box were, height of the box, width of the box, length of the box, and Window to Wall Ratio (WWR) of the glazing provided just on the southern façade. The main feature used for daylight analysis was <strong>“Daylight Autonomy”</strong>.</p><blockquote><strong>Daylight autonomy</strong> is the percentage of time that <strong>daylight</strong> levels are above a specified target illuminance within a physical space or building. The calculation is based on annual data and the predetermined lighting levels. — <a href="https://en.wikipedia.org/wiki/Daylighting#:~:text=Daylight%20autonomy%20is%20the%20percentage,and%20the%20predetermined%20lighting%20levels.">Wikipedia</a></blockquote><p>The city chosen for this analysis was the Indian city of Mumbai, which is a moderately hot tropical location.</p><p>The daylight simulations were computationally intensive, so input parameters were limited to certain values. In our experiment, a total of 117 design options and simulation daylight simulations were created. The dataset contained geometric and glazing information for each design option.</p><h3>Data Analysis</h3><p>Here is the correlation of all the features with Daylight Autonomy (DA):</p><p><em>1. Volume and Average DA = .22</em></p><p><em>2. Floor Area and Average DA = -0.04</em></p><p><em>3. Width and Average DA = .23</em></p><p><em>4. Length and Average DA = -0.48 (Strong negative correlation)</em></p><p><em>5. Height and Average DA = 0.53 (Strong positive correlation)</em></p><p><em>6. WWR and Average DA = 0.34</em></p><p>When two sets of data are strongly linked together we say they have a High Correlation. Correlation is Positive when the values increase together, and. Correlation is Negative when one value decreases as the other increases.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/816/1*7HPOerRmNR5OULhUIxftbw.png" /><figcaption>The length of the room is inversely proportional to Daylight Autonomy. A threshold length of 8 meters is considered the maximum for a single glazing room.</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/816/1*EBme9a9_kq8dEzciIiJa-Q.png" /><figcaption>This graph shows, there was hardly any correlation of floor area with daylight autonomy, but room height and Window to wall ratio had a huge impact on average daylight autonomy (DA).</figcaption></figure><h3>Results of Machine Learning</h3><p>The daylighting data was split into two sub-datasets: train data, and test data. The simple linear regression model was trained on the train data with the length, width, height, WWR, floor area as the model features, and Daylight Autonomy, as the target variable. The Daylight Autonomy was predicted for the test data using the trained ML model.</p><p><em>The machine learning process is mentioned below.</em></p><blockquote><strong>Process → Split → Train → Predict → Evaluate</strong></blockquote><p>Just by using simple linear regression to predict the Daylight Autonomy, the model was able to achieve a correlation of 96.7% , which is actually great. more features can be added to this like the location, glazing on other façades, and much more. This can be a more data-driven approach to sustainable designing and can even cut the time taken for heavy daylight simulations.</p><blockquote>“Designers should be able to use sunlighting strategies without creating a conspicuous ‘sunlighting building’ look.” — William Lam</blockquote><p><strong>N O T E S</strong></p><p>1 “Daylighting Design Principles” <a href="https://ocw.mit.edu/courses/architecture/4-401-environmental-technologies-in-buildings-fall-2018/lecture-slides-1/MIT4_401F18_lec8.pdf">https://ocw.mit.edu/courses/architecture/4-401-environmental-technologies-in-buildings-fall-2018/lecture-slides-1/MIT4_401F18_lec8.pdf</a></p><p>2 “Daylight Rules-of-Thumb Experimentally Examined” <a href="https://www.asee.org/documents/sections/2017/Mid-West/32-Mansy-Daylight%20Rules-of-Thumb%20Experimentally%20Examined.pdf">https://www.asee.org/documents/sections/2017/Mid-West/32-Mansy-Daylight%20Rules-of-Thumb%20Experimentally%20Examined.pdf</a></p><p>3 “Daylighting Buildings” ; <a href="http://www.ewinlight.com/wp-content/uploads/2017/09/grtdaylightingpresn2009-11-12-100730124931-phpapp01.pdf">http://www.ewinlight.com/wp-content/uploads/2017/09/grtdaylightingpresn2009-11-12-100730124931-phpapp01.pdf</a></p><p>4 “Daylight design rules of thumb” <a href="https://www.irbnet.de/daten/iconda/CIB_DC23487.pdf">https://www.irbnet.de/daten/iconda/CIB_DC23487.pdf</a></p><h4>About the Author</h4><p><strong>Rutvik Deshpande</strong> is a machine learning and data science research intern at<a href="https://www.digitalbluefoam.com/"> Digital Blue Foam</a> and Architecture Student at<a href="https://twitter.com/NITRR"> @NITRR</a>. He has interests in Data Science, Machine Learning, and Artificial Intelligence for the AEC industry. He is also a Computational Design, Urban Analytics enthusiast and a<a href="https://www.kaggle.com/rutvikdeshpande"> Kaggler</a> too.</p><h4>About Digital Blue Foam</h4><p>At <a href="https://www.digitalbluefoam.com/">Digital Blue Foam</a> we believe buildings and cities must reduce their dependence on fossil fuels in all aspects of the process design. This includes materials, construction processes, and operations. Presently, we are researching new ways to reduce energy use , and ultimately carbon emissions, and boost occupant comfort by leveraging tried and tested passive design principles in our design software.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=412166369a3d" width="1" height="1" alt=""><hr><p><a href="https://medium.com/digital-blue-foam/using-artificial-intelligence-to-predict-building-daylight-autonomy-412166369a3d">Using Artificial Intelligence to Predict Building Daylight Autonomy</a> was originally published in <a href="https://medium.com/digital-blue-foam">Digital Blue Foam</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Interactive AI for Sustainable Architecture]]></title>
            <link>https://medium.com/digital-blue-foam/interactive-ai-for-sustainable-architecture-c5baf4c0ad3d?source=rss-7970904ad169------2</link>
            <guid isPermaLink="false">https://medium.com/p/c5baf4c0ad3d</guid>
            <category><![CDATA[architecture]]></category>
            <category><![CDATA[generative-design]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[sustainable-development]]></category>
            <category><![CDATA[analysis]]></category>
            <dc:creator><![CDATA[Rutvik Deshpande]]></dc:creator>
            <pubDate>Thu, 27 Aug 2020 02:22:24 GMT</pubDate>
            <atom:updated>2020-08-27T02:22:24.496Z</atom:updated>
            <content:encoded><![CDATA[<h4>Augmented Intelligence</h4><h4>Digital Blue Foam</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*BPHSBBOxLY4OyMa72PZjeA.png" /><figcaption><a href="https://www.digitalbluefoam.com/">Digital Blue Foam</a></figcaption></figure><p>I was recently selected for a research internship at one of the most promising and exciting AEC startups — <a href="https://www.digitalbluefoam.com/">Digital Blue Foam</a>. I was given a chance to use their Augmented Intelligence enhanced software. Here’s an article that describes its features and advantages.</p><p>So this software is a <strong>web-based interactive generative design tool</strong> that gathers urban and climate data from various available web sources. Once you log into the Digital Blue Foam website you get options to load the previous session (not for you if it’s your first time), search bar for searching and locating the site you are looking for designing, and feeling lucky option, which takes you through the tool and guides you about its features (you must select this option if you want to learn about the tool in brief).</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/1*u-kRlVzt5LxTkiVoPbjk4A.gif" /></figure><h4><strong>Locate and Select Site</strong></h4><p>Once you search for the location the tool will fetch various datasets and take you to the location where you can draw and edit a polygon. Now you can see your located site in the 3D surrounding map.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/1*407M9RAvwimKLdV5kq1cBQ.gif" /><figcaption>You can draw a polygon to select your site</figcaption></figure><h4><strong>Draw Lines using the Pencil and Eraser</strong></h4><p>You can draw a random line or a polyline between your plot to divide it into subplots. The pencil has three options: (1) Axis — to assign the axis of the plot; (2)- Park, which creates and assigns the drawn place to park; and (3) a tape measure tool for measuring the line you are drawing.</p><p>Upon double-clicking, the click and drag draw a line option the line for subplots is finalized, which can be seen in orange color. Being an interactive software the number of subplots can even be changed in the statistics section, about which I’ll inform you in this article.</p><p>Draw tool also has an erase feature in case you want to redraw the lines. Clicking the play button now will generate designs according to the subplots drawn.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/1*AumBqfod2o9TS_dgsjdXQg.gif" /><figcaption>Straight as well as curved lines can be drawn to divide the plot</figcaption></figure><h4>Generate Designs</h4><p>The play button generates designs. You can pause on the design you like. The step forward and step backward are buttons that will toggle the designs one at a time.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/1*1RXGxYOWnv3XhOQ0N9PvWQ.gif" /><figcaption>The play button generates many options for the user to choose from.</figcaption></figure><h4><strong>Program Distribution</strong></h4><p>This section has input sliders that can be changed to change the ratio of educational, residential, commercial, office, and leisure spaces according to your needs.</p><p>If you like the generated design and not the distribution of program spaces you can click on the specific part of the structure which gets highlighted and then you can edit features like the program, floor height, and the number of floors in DBF Assistant. The side arrow keys give you the freedom to move the selected block. Selecting the plots also offers various options like convert the plot to park, parking, or a podium. You can even merge 2 or more plots.</p><h4><strong>Sustainable Tools</strong></h4><p>This tool has 3 features:</p><p><strong>a) Sun path</strong></p><p>This feature helps in the visualization of sun paths and shadows for your site. It also has input date and time for users to select. While designing a building, checking the sun path and shadows is essential as studies have shown that sufficient daylight in space increases one’s productivity and freshness quotient.</p><p><strong>b) Wind path</strong></p><p>This feature allows you to visualize the direction of the wind with huge arrows.</p><p>Wind path is another important factor in sustainable and climate-responsive design. The location of the opening can be chosen according to the wind direction. Also, it helps to analyze the airflow of the plot.</p><p><strong>c) Solar radiation</strong></p><p>This is the third sustainability tool. The Solar Radiation option does make the designing process a bit slow, but of course, this option helps you to visualize and analyze the distribution of solar radiation throughout your design. It can be beneficial while designing the shading requirements for the structure.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/1*O_hIDO9aYWon3bU_1xftDg.gif" /><figcaption>Sun path and Wind Direction can be selected</figcaption></figure><h4><strong>Statistics Section</strong></h4><p>This section offers a wide range of targets to choose from. Max Height., Facade Area, Floor Area, Site Area, and other targets can be set by the user. One of my favorite targets among these options is Site Efficiency, which tells you the site efficiency percentage. You can see this target change by changing the subplots and programs. This makes it an interactive and informative tool that has sustainability options to choose from.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/1*Iukk-vAN4VZDvLxc-npNJQ.gif" /><figcaption>Multiple input targets that can be set for the design, which includes Efficiency, GFA, and many more</figcaption></figure><h4><strong>Viewing Tool</strong></h4><p>This tool has 3 subtypes:-</p><p><strong>a) View</strong> — This consists of 4 features:</p><ul><li>Orbital view</li><li>Toggle view</li><li>Option to show cars</li><li>Show trees in the plot</li></ul><p><strong>b) Plan view</strong> — This part also has 4 features:</p><ul><li>Plan view</li><li>Longitudinal section</li><li>Cross-section</li><li>Toggle isometric view</li></ul><p><em>All these views are important from an architectural viewpoint.</em></p><p><strong>c) Visualize Data</strong> — this subtype has 5 features:</p><ul><li>Terrain — to visualize the satellite imagery of the map</li><li>Neighborhood programs — Proximity analysis of various nearby facilities</li><li>Heatmap</li><li>Structural Details</li></ul><h4><strong>Save and Download</strong></h4><p>a) You will have slots to save 5 of your designs. These saved designs can be compared using the compare design tool, this compares factors like FAR (Floor Area Ratio), GFA (Gross Floor Area), and Efficiency.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*k8PtO99QvxhLU2C6MzJvqg.png" /><figcaption>An insightful comparison of up to 5 designs</figcaption></figure><p>b) The download button offers 4 options consisting of screenshot, 3D, 2D, and download an excel (.xlsx file) report. The report shows details like floor height, and area for every floor in each program. This file also has input sections like Title, Building Type, Site Area, which can be useful for presentations and documentation.</p><p>What I liked most about this tool is the designs are generated on clicking the play button and you can see the targets change their values. When you feel the right target has been achieved then you can pause and save the design. This tool is one of the best-use cases of <strong>“AI in AEC”</strong>, which helps the user in the optimization of the design efficiently with several input options.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/1*PJK6Veg24KBYGoNZUcIqog.gif" /><figcaption><a href="https://www.digitalbluefoam.com/">Digital Blue Foam</a></figcaption></figure><p>The user interface is easy to use. Having a <em>DBF Assistant</em> makes it simple for the user to get information about the various tools offered. The software also has options like Quick Tips and Tutorials to guide you through the tool.</p><p>Progress is impossible without a change and it has become an important factor to change our designing process, We should make it more efficient and sustainable. The way your city looks today will be very different in the coming years, So it is important to optimize your design and analyze its effects on the environment for future generations.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/939/1*qxj0a1F76Sy5cSoE0aygzg.png" /></figure><blockquote>“The future of urban design and project development has arrived” — Digital Blue Foam</blockquote><p>Visit <a href="http://digitalbluefoam.com/">digitalbluefoam.com</a> to learn about Digital Blue Foam’s early access program!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c5baf4c0ad3d" width="1" height="1" alt=""><hr><p><a href="https://medium.com/digital-blue-foam/interactive-ai-for-sustainable-architecture-c5baf4c0ad3d">Interactive AI for Sustainable Architecture</a> was originally published in <a href="https://medium.com/digital-blue-foam">Digital Blue Foam</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How to Choose the Ideal Site for Designing Your Restaurant Using Data Science]]></title>
            <link>https://medium.com/swlh/how-to-choose-the-ideal-site-for-designing-your-restaurant-using-data-science-2cbfb9853f93?source=rss-7970904ad169------2</link>
            <guid isPermaLink="false">https://medium.com/p/2cbfb9853f93</guid>
            <category><![CDATA[data-visualization]]></category>
            <category><![CDATA[restaurant]]></category>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[analytics]]></category>
            <category><![CDATA[data-science]]></category>
            <dc:creator><![CDATA[Rutvik Deshpande]]></dc:creator>
            <pubDate>Mon, 10 Aug 2020 15:01:32 GMT</pubDate>
            <atom:updated>2020-08-10T15:01:32.258Z</atom:updated>
            <content:encoded><![CDATA[<h4>Spatial Data Science</h4><h4>Designing any product requires a lot of analysis and research. Restaurants are no different.</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*-FmljN2xwuOfcjGd4hTI1Q.jpeg" /><figcaption>Photo by <a href="https://unsplash.com/@terminath0r?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Thor Alvis</a> on <a href="https://unsplash.com/s/photos/restaurant-and-map?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Unsplash</a></figcaption></figure><p>Designing any product requires a lot of analysis and research. It is also true for designing any building. Before we begin to design any building, we collect information about the location where we are designing, we check the budget, we do research about the background of the target audience and their likes and dislikes.</p><p>In designing any space you need to follow some preliminary steps, like budget analysis, client requirements, form &amp; concept development, site analysis, and zoning.</p><p>Site Analysis is one of the major steps in the pre-designing processes. It involves analysis and evaluating topography, watercourses, trees, manmade features, site boundaries, habitats, wind circulations, connectivity to main roads, streets and pathways, distance to closest facilities or amenities, climatic conditions, and weather patterns.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1000/1*tVNgPQsSUswYOsUUOYSJYQ.jpeg" /><figcaption>Photo Credit: <a href="https://www.seradesign.com/">SERA Architects</a></figcaption></figure><p>In my <strong>Architectural Design</strong> course, we were given the assignment to design a restaurant in our home town, with very minimum and basic specifications. Having complete freedom to choose the site, concept, form, building materials, and cuisine offered, it required a lot of thinking and creativity.</p><p>But it is very difficult to locate the ideal location in such a big city as it also requires a huge amount of research to locate a site because the restaurant should also be successful in business after it starts to function. So instead of choosing my favorite or city’s most popular restaurant’s or cafe’s location as my site, I thought of implementing Data Science for analyzing the data of the restaurants in Pune (my home city, in the western Indian state of Maharashtra) and then using Machine Learning to locate the ideal site.</p><p>For Data Scienc<em>e</em> you need a large amount of data for the results to be precise, so I collected data about restaurants in Pune through Zomato API’s, population data, pollution data, and keeping in mind the current situation, I even used the geospatial data of COVID-19 cases, from various available sources.</p><p>But the COVID-19 data was of not that use as the number of cases and containment zones change every day. So whatever area the Restaurant is in, precautions should be taken.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*jkRGeiIXO7WyxNA5" /><figcaption>Photo by <a href="https://unsplash.com/@clintadair?utm_source=medium&amp;utm_medium=referral">Clint Adair</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><h4>For the analysis, I used the following factors to locate the ideal location in the city:</h4><ol><li><em>Ratings</em></li><li><em>Votes</em></li><li><em>Geolocation of the Restaurant</em></li><li><em>Locality / Neighbourhood</em></li><li><em>Price Per Sq. Feet of the Commercial plot of each Neighbourhood</em></li><li><em>Cost for two people</em></li><li><em>Type of the Restaurant</em></li></ol><p>I observed and analyzed each and every factor for location analysis and prediction, and only considered the factors that were important for me. There can be many more factors but due to data availability constraints, I used the above-mentioned factors only.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*TILz29rDLuL3c7El" /><figcaption>Photo by <a href="https://unsplash.com/@krisatomic?utm_source=medium&amp;utm_medium=referral">Kris Atomic</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p>Locating the ideal site is important but it is not the only factor to be considered in the planning phase, you should also decide what you are going to serve, price of the land or the rent and what should be the ideal cost. These factors should also be analyzed wisely as only locating the site will not get you customers and rating it is also important where you serve what and for what price.</p><p>For example, if the ideal location based on your budget is the “Business and Office” area then Casual Dining would be an appropriate Type and the cost for two people can be more so you can invest more things like formal furniture and menu, if you are planning for a restaurant in an area which has many schools, colleges, and coaching classes then Cafe might be the best alternative as students prefer cheap food and a good place for hanging out. This is just a rough idea or an example, before actual analysis.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/1*QShssCioBy3cD6ZLGwACUA.png" /></figure><p>Popularity and Rating are two different factors for judging a Restaurant, popularity can be seen through the number of votes. When it comes to popular localities Kothrud is the most popular locality for cafes, restaurants, and eateries as it is a densely populated area with several colleges and schools. Kothrud is followed by Viman Nagar and then Hinjawadi. Hinjawadi is a corporate location consisting of a high number of offices.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/830/1*8fvWP5WG-CjxnjNCTZuIvw.png" /></figure><p>Locality or neighborhood or suburb are all broader terms. This is the list of top 10 specific locations that serve good food as they have a rating of above 4. Rating symbolizes customer satisfaction whereas the number of votes tells us the average footfall of that restaurant.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/1*QfD5jEU2CDArfGokPdrARw.png" /></figure><p>This graph tells the average rating of each locality and is ranked accordingly. Kothrud is a popular eatery hub, but in terms of great food and good service, Baner occupies the top position. This can be a crucial factor while deciding the location for a restaurant.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/1*HI14YoOGbj69IeFamNkLvQ.png" /></figure><p>Not every type of restaurant works everywhere. Popularity and Rating are major factors for deciding the ideal location but the type of restaurant also plays a crucial role in success. Kothrud might be the most popular locality in Pune but not necessarily every restaurant type will work there. For example, Casual Dining has the most number of votes in Viman Nagar, Dessert Parlour, or Bar has popularity in Baner. The popularity of a type in a specific location depends on the type of audience living there. Baner and Aundh are popular hangout places for a younger generation so Microbrewery, Pub, and Dessert Parlour gain more audience here.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/1*azZ4gu1KRTxiUmxl2PHtaw.png" /></figure><p>The cost of the food never decides the popularity or rating of a restaurant. But this graph shows an ideal range price depending on the rating. For the rating to be 4.0 &lt; the ideal cost for two people should range from ₹600 — ₹1200. No restaurant in Pune was ever rated 5.0.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/910/1*r_NhRGq0eKjEykVpiTNAkg.png" /></figure><p>When locating an ideal restaurant site is done the next time aspect for consideration is the budget. The price of commercial lands is not even close to similar. <strong><em>Wakad</em></strong> has the costliest commercial land, followed by <strong><em>Balewadi</em></strong>. The price of the land does not play any role in deciding the success of the restaurant but budget planning is also important while choosing the location.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/900/1*laMh8JGqJ407OEIeN5yURA.png" /></figure><p>This gives a general idea of which type of restaurants are popular in Pune. Quick Bites and Casual Dining are popular in Pune, the city being an educational hub as well as the IT hub.</p><p>I trained this data in one of the Machine Learning algorithms and deployed the web application. Users can choose the input values as per their choice and the ideal locality based on the input values is predicted. <em>(I have deployed the web application using shiny.io)</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*8Q7JHSoa-_Trn1Ds849f3A.png" /></figure><blockquote><a href="https://localitypredictor.shinyapps.io/restolocator/">https://localitypredictor.shinyapps.io/restolocator/</a></blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*nb89duyZsesC4M3_ta9gLA.png" /></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=2cbfb9853f93" width="1" height="1" alt=""><hr><p><a href="https://medium.com/swlh/how-to-choose-the-ideal-site-for-designing-your-restaurant-using-data-science-2cbfb9853f93">How to Choose the Ideal Site for Designing Your Restaurant Using Data Science</a> was originally published in <a href="https://medium.com/swlh">The Startup</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Predicting Bike Sharing Demand using RandomForest Algorithm]]></title>
            <link>https://medium.com/analytics-vidhya/predicting-bike-sharing-demand-using-randomforest-algorithm-32682a7a07e7?source=rss-7970904ad169------2</link>
            <guid isPermaLink="false">https://medium.com/p/32682a7a07e7</guid>
            <category><![CDATA[kaggle]]></category>
            <category><![CDATA[competition]]></category>
            <category><![CDATA[predictive-analytics]]></category>
            <category><![CDATA[beginner]]></category>
            <category><![CDATA[data-science]]></category>
            <dc:creator><![CDATA[Rutvik Deshpande]]></dc:creator>
            <pubDate>Sun, 12 Jul 2020 13:04:29 GMT</pubDate>
            <atom:updated>2020-07-16T15:45:39.384Z</atom:updated>
            <content:encoded><![CDATA[<h4>In-Depth Analysis</h4><h4>A Beginners Guide to Kaggle Competition</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*fd8xKJrry0QRlqINDAzEsQ.jpeg" /><figcaption>Photo by <a href="https://unsplash.com/@jorgeaalcala?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Jorge Alcala</a> on <a href="https://unsplash.com/s/photos/washington-d.c?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Unsplash</a></figcaption></figure><h3>Introduction</h3><p>As urbanisation have reached unprecedented levels, road congestion has become a modern day issue. Heavy traffic is associated with air pollution, safety risks, economic competitiveness, sustainable growth and social cohesion.</p><p>In this situation<em> Bike sharing systems</em> are becoming a major solution to avoid the above mentioned issues. The benefits of bike sharing schemes include transport flexibility, reductions to vehicle pollutants, health benefits, reduced congestion and fuel consumption, and financial savings for individuals.</p><p>This is a report on Data Science for beginners and enthusiasts starting Kaggle competitions. The following is detailed description of <em>Data Science</em> competition which includes <em>Data Manipulation</em>,<em> Data Visualization</em> and using <em>Machine Learning Algorithms</em> for predicting the <a href="https://www.kaggle.com/c/bike-sharing-demand"><strong>Bike Sharing Demand</strong></a> in the city of Washington D.C., also it gives step by step guide for Prediction of sharing demand.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/613/1*U2ZINEkrGWkZ0tfV5EhskQ.png" /></figure><pre>## Loading the required packages</pre><pre>library(tidyverse)<br>library(dplyr)<br>library(scales)<br>library(ggplot2)<br>library(glmnet)<br>library(caret)<br>library(lubridate)</pre><p>I used popular Machine Learning Meta Algorithm RandomForest, which an ensemble learning method for classification and regression. It is also one of the most used algorithms, because it produces great without hyperparameter tuning.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*EagtUIn7gSonJpAO" /><figcaption>Photo by <a href="https://unsplash.com/@sebastian_unrau?utm_source=medium&amp;utm_medium=referral">Sebastian Unrau</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><pre>## Importing the data <br>biketrain &lt;- read.csv(&quot;../input/bike-sharing-demand/train.csv&quot;)<br>biketest &lt;- read.csv(&quot;../input/bike-sharing-demand/test.csv&quot;)</pre><pre>## Making equal no. of columns in boths datasets <br>biketest$registered=0<br>biketest$casual=0<br>biketest$count=0</pre><pre>## Joining both train and test datasets</pre><pre>bikesharing &lt;- rbind(biketrain, biketest)</pre><pre>## Inspecting the data</pre><pre>summary(bikesharing)</pre><pre>## season -  1 = spring, 2 = summer, 3 = fall, 4 = winter <br>## weather - 1: Clear, Few clouds, Partly cloudy, Partly cloudy ; 2: Mist + Cloudy, Mist + Broken clouds, Mist + Few clouds, Mist ;3: Light Snow, Light Rain + Thunderstorm + Scattered clouds, Light Rain + Scattered clouds ;4: Heavy Rain + Ice Pallets + Thunderstorm + Mist, Snow + Fog</pre><pre>## Converting a few a variables into factors<br>bikesharing$season &lt;- as.factor(bikesharing$season)<br>bikesharing$weather &lt;- as.factor(bikesharing$weather)<br>bikesharing$holiday &lt;- as.factor(bikesharing$holiday)</pre><h3><strong>Feature Engineering</strong></h3><p>Lets work on the first variable <em>datetime</em> , because we can break it down into many additional meaningful variables like Hour of the Day , Day of the Month , Day of the Week , Hour of the Day , Month , Year , which can useful for further Data Analysis and we can feed them for Predicting sharing demands .</p><p><strong>1. Hour</strong></p><pre>## Extracting hour from date time</pre><pre>bikesharing$hour = substr(bikesharing$datetime,12,13)<br>bikesharing$hour &lt;- as.factor(bikesharing$hour)</pre><pre>## plotting the graph</pre><pre>ggplot(data=bikesharing, aes(x = hour, y=count, fill=hour))+geom_bar(stat = &quot;identity&quot;)+ggtitle(label = &quot;Hourly Comparison with Seasons&quot;)+theme_minimal()+theme(plot.title = element_text(hjust = 0.5, lineheight = 0.8, face = &quot;bold&quot;),legend.position = &quot;none&quot;)+xlab(&quot;Weekdays&quot;)+ylab(&quot;Count&quot;)</pre><figure><img alt="" src="https://cdn-images-1.medium.com/max/764/1*XtqUjZtV-Jyd8pgiwK9wEQ.png" /></figure><p>People in the city used the bike maximum in between 5–8 pm , also there was a surge in demand between 8–9 am during office hours, to maybe avoid traffic.</p><pre>## Season and Hour vs Count Graph</pre><pre>ggplot(bikesharing, aes(hour, count)) + <br>  geom_bar(aes(fill = season), position = &#39;dodge&#39;, stat = &#39;identity&#39;) + <br>  ggtitle(label=&quot;Count with Hours of Day and Seasons&quot;) + scale_y_continuous(labels = comma) + theme(plot.title = element_text(hjust = 0.5, lineheight = 0.8, face = &quot;bold&quot;)) + theme_minimal()</pre><figure><img alt="" src="https://cdn-images-1.medium.com/max/729/1*rcGQcVCt34Zd79CrMnKPJQ.png" /></figure><p>Fall and Winter were main seasons where demand was highest across most of the hours of day . According to cycling experts , Fall Season is considered the best for cycling as , you dont get all sweaty after the ride , colour of the leaves keep on changing and it feels pleasant to ride a bike.</p><blockquote><em>“Spring passes and one remembers one’s innocence ,</em></blockquote><blockquote><em>Summer passes and one remembers one’s exuberance ,</em></blockquote><blockquote><em>Fall passes and one remembers one’s reverence ,</em></blockquote><blockquote><em>Winter passes and one remembers one’s perseverance”</em></blockquote><h4><strong>2. Day</strong></h4><pre>## Daylight</pre><pre>bikesharing$light &lt;- as.factor(bikesharing$hour == c(7:19))<br></pre><pre>## Holiday</pre><pre>bikesharing$holiday &lt;- as.factor(bikesharing$holiday)</pre><pre>## Weeked or not</pre><pre>bikesharing$weekend=0<br>bikesharing$weekend[bikesharing$day==&quot;Sunday&quot; | bikesharing$day==&quot;Saturday&quot; ]=1</pre><pre>## Free time in a week</pre><pre>bikesharing$officehour &lt;- as.factor(bikesharing$hour == c(9:17))<br>bikesharing$freehourworkday &lt;- as.factor(bikesharing$officehour == FALSE &amp;&amp; bikesharing$holiday == 0)</pre><pre>## Day of the Week</pre><pre>bikesharing$date = substr(bikesharing$datetime,1,10)<br>days&lt;-weekdays(as.Date(bikesharing$date))<br>bikesharing$days=days</pre><pre>bikesharing$days &lt;- factor(bikesharing$days, levels=c(&quot;Monday&quot;,&quot;Tuesday&quot;, &quot;Wednesday&quot;, &quot;Thursday&quot;, &quot;Friday&quot;, &quot;Saturday&quot;,&quot;Sunday&quot;))</pre><pre>## Weekdays Graph</pre><pre>ggplot(data=bikesharing, aes(x = days, y=casual, fill=days))+geom_bar(stat = &quot;identity&quot;)+ggtitle(label = &quot;Weekday Comparison of Casual Users&quot;)+theme_minimal()+theme(plot.title = element_text(hjust = 0.5, lineheight = 0.8, face = &quot;bold&quot;),legend.position = &quot;none&quot;)+xlab(&quot;Weekdays&quot;)+ylab(&quot;Casual Count&quot;)+ scale_y_continuous(labels = comma)</pre><figure><img alt="" src="https://cdn-images-1.medium.com/max/729/1*LapAsvIggAkLt0rSYrX2bA.png" /></figure><p>Casual users were more on weekends that is , they did not require a registration as they rode the bikes in their free time only . These maybe onetime users .</p><pre>## Registered Users and Count across Weekdays</pre><pre>ggplot(data=bikesharing, aes(x = days, y=registered, fill=days))+geom_bar(stat = &quot;identity&quot;)+ggtitle(label = &quot;Weekday Comparison of Registered Users&quot;)+theme_minimal()+theme(plot.title = element_text(hjust = 0.5, lineheight = 0.8, face = &quot;bold&quot;),legend.position = &quot;none&quot;)+xlab(&quot;Weekdays&quot;)+ylab(&quot;Registered Count&quot;) + scale_y_continuous(labels = comma)</pre><figure><img alt="" src="https://cdn-images-1.medium.com/max/729/1*ycdtmzqdW4XQqhwxxCsNoA.png" /></figure><p>Registered no. of users peaked on weekdays , using the bikes to avoid traffic and reaching office on time , so they must have registered as they require bikes almost everyday .</p><pre>ggplot(data=bikesharing, aes(x = days, y=count, fill=days))+geom_bar(stat = &quot;identity&quot;)+ggtitle(label = &quot;Weekday Comparison&quot;)+theme_minimal()+theme(plot.title = element_text(hjust = 0.5, lineheight = 0.8, face = &quot;bold&quot;),legend.position = &quot;none&quot;)+xlab(&quot;&quot;)+ylab(&quot;Total Count&quot;) + scale_y_continuous(labels = comma)</pre><figure><img alt="" src="https://cdn-images-1.medium.com/max/729/1*TscPLP91Qgzfo0K6nYTnsg.png" /></figure><pre>## Weekday and Weather Comparison</pre><pre>ggplot(data=bikesharing, aes(x=days, y=count, fill=weather)) +<br>  geom_bar(stat=&quot;identity&quot;, position=position_dodge()) + ggtitle(label =&quot;Day of the Week and Weather with Count Comparison&quot;) +theme_minimal()+theme(plot.title = element_text(hjust = 0.5, lineheight = 0.8, face = &quot;bold&quot;))+xlab(&quot;&quot;)+ylab(&quot;&quot;)</pre><figure><img alt="" src="https://cdn-images-1.medium.com/max/729/1*6IWCwvakYTegqontJ4j-QQ.png" /></figure><p><strong>3. Month</strong></p><pre>## Converting date to a date format in R</pre><pre>bikesharing$date &lt;- as.POSIXct(bikesharing$date)</pre><pre>## Extracting Month Name</pre><pre>bikesharing$MONTHS &lt;- format(bikesharing$date,&quot;%B&quot;)</pre><pre>bikesharing$MONTHS &lt;- factor(bikesharing$MONTHS, levels=c(&quot;January&quot;,&quot;February&quot;, &quot;March&quot;, &quot;April&quot;, &quot;May&quot;, &quot;June&quot;,&quot;July&quot;, &quot;August&quot;, &quot;September&quot;, &quot;October&quot;, &quot;November&quot;, &quot;December&quot;))</pre><pre>## Monthly Graph</pre><pre>ggplot(data=bikesharing, aes(x = MONTHS, y=count, fill=MONTHS))+geom_bar(stat = &quot;identity&quot;)+ggtitle(label = &quot;Monthly Comparison of Users&quot;)+theme_minimal()+theme(plot.title = element_text(hjust = 0.5, lineheight = 0.8, face = &quot;bold&quot;), legend.position = &quot;none&quot;)+xlab(&quot;&quot;)+ylab(&quot;Count&quot;)</pre><figure><img alt="" src="https://cdn-images-1.medium.com/max/729/1*W0r0Jdp8rzy-MoWIVlq5Fg.png" /></figure><p>Second half of the year may have been good to ride the bike as we can see , peak in demand can be seen in the fall season .</p><pre>## Month and Weekday vs Count</pre><pre>ggplot(data=bikesharing, aes(x = MONTHS, y=count, fill=days))+geom_bar(stat = &quot;identity&quot;)+ggtitle(label = &quot;Users Count compared in Months and Weekdays&quot;)+theme_minimal()+theme(plot.title = element_text(hjust = 0.5, lineheight = 0.8, face = &quot;bold&quot;))+xlab(&quot;&quot;)+ylab(&quot;&quot;)</pre><figure><img alt="" src="https://cdn-images-1.medium.com/max/729/1*4j6GJw3-JEQJMIeuzv6i2A.png" /></figure><pre>## Day of the Month Extraction</pre><pre>bikesharing$dom = substr(bikesharing$datetime,9,10)<br>bikesharing$dom &lt;- as.numeric(bikesharing$dom)</pre><pre>## Season and Day of Month vs Count of Users</pre><pre>ggplot(data=bikesharing, aes(x = dom, y=count, fill=season))+geom_bar(stat = &quot;identity&quot;)+ggtitle(label = &quot;Day of the Month with Season Comparison&quot;)+theme_minimal()+theme(plot.title = element_text(hjust = 0.5, lineheight = 0.8, face = &quot;bold&quot;))+xlab(&quot;Days upto 19th&quot;)+ylab(&quot;Count&quot;) + xlim(01,19)</pre><figure><img alt="" src="https://cdn-images-1.medium.com/max/729/1*gdHyto9m3TgRxAPzIuEwpg.png" /></figure><pre><em>## Visualizing the demand only till 19 th of every Month as we have to predict the number of users from 20th to end of each month</em></pre><pre>## Plotting a Heat Tile Graph</pre><pre>ggplot(bikesharing, aes(dom, hour, fill = count)) + geom_tile(color = &quot;white&quot;) +ggtitle(label = &quot;Day of the Month and Hour of the Day Comparison&quot;)+theme_minimal()+theme(plot.title = element_text(hjust = 0.5, lineheight = 0.8, face = &quot;bold&quot;)) + xlab(&quot;Day of the Month(Till 19th)&quot;) + ylab(&quot;Hour&quot;)</pre><figure><img alt="" src="https://cdn-images-1.medium.com/max/729/1*kgtjus1JoALnfuVqITfx1w.png" /></figure><p><strong>We can Treat a few more variables to extract some more additional information</strong></p><p>What else we can do here ? Useful information can be extracted from humidity, temperature, windspeed, weekdays, lets check it out.</p><p><strong>4. Weather Conditions</strong></p><p>Best conditions to ride are when you have sunshine that creates great visibility, less windspeed, less humidity to avoid sweating during high temperatures.</p><pre>## Ideal Condition to ride a bike</pre><pre>bikesharing$Ideal &lt;- as.factor(bikesharing$weather == 1 &amp;&amp; bikesharing$windspeed &lt; 20 &amp;&amp; bikesharing$humidity &lt; 25 &amp;&amp; bikesharing$atemp &lt; 35)</pre><pre>## Calculation Pollution(Ozone Content) using Temp, Windspeed and Humidity</pre><pre>bikesharing$ozone &lt;- 5.3*(bikesharing$windspeed)+0.4*(bikesharing$temp)+0.1*(bikesharing$humidity)</pre><pre>## Plotting Weather and Temp vs Count Graph</pre><pre>ggplot(bikesharing, aes(temp, count)) + geom_jitter(aes(colour = weather)) + ggtitle(label =&quot;Effect of Weather and Temperature on Count&quot;)+theme_minimal()+theme(plot.title = element_text(hjust = 0.5, lineheight = 0.8, face = &quot;bold&quot;)) + xlab(&quot;Temperature&quot;) + ylab(&quot;Count of Users&quot;)</pre><figure><img alt="" src="https://cdn-images-1.medium.com/max/729/1*RrWJ9jvcyWSObDVxI9MAUQ.png" /></figure><pre>## Windspeed &amp; Hour of the day vs Users&#39; Count</pre><pre>ggplot(bikesharing, aes(windspeed, hour)) + geom_jitter(color= &quot;blue&quot;) + theme_minimal() + ggtitle(label = &quot;Windspeed and Hour of the Day with Users&#39; Count&quot;)+theme_minimal()+theme(plot.title = element_text(hjust = 0.5, lineheight = 0.8, face = &quot;bold&quot;)) + xlab(&quot;Windspeed&quot;) + ylab(&quot;Hour&quot;)</pre><figure><img alt="" src="https://cdn-images-1.medium.com/max/729/1*wnmwhFI-86FUcckv02e3hQ.png" /></figure><p>Create new log variables</p><ul><li>As we know that dependent variables have natural outliers so we will predict log of dependent variables.</li><li>Predict bike demand of registered and casual users separately, logcas=log(casual+1) and logreg=log(registered+1),add 1 to deal with zero values .</li></ul><pre>bikesharing$logcas &lt;- log(bikesharing$casual + 1)<br>bikesharing$logreg &lt;- log(bikesharing$registered + 1)</pre><h3>Model Deployment and Predcition</h3><pre>## Converting numeric values to factor for randomForest</pre><pre>bikesharing = bikesharing %&gt;% mutate_if(is.factor, as.numeric)</pre><pre>str(bikesharing)</pre><pre>## Train and Test data separation</pre><pre>biketrain &lt;- bikesharing[1:10886, ]<br>biketest &lt;- bikesharing[10887:17379, ]</pre><pre>## Deploy the model</pre><pre>set.seed(8234)<br>library(randomForest)</pre><pre>## Casual</pre><pre>bkmodel &lt;- randomForest(logcas ~ season + holiday + workingday + weather + temp + atemp + humidity + windspeed + Ideal + hour + MONTHS + days + officehour + freehourworkday + weekend + ozone + light + dom, data = biketrain)</pre><pre># Predict</pre><pre>predict1 &lt;- predict(bkmodel, biketest)<br>summary(bkmodel)</pre><pre>## Registered </pre><pre>set.seed(9234)<br>bkmodel1 &lt;- randomForest(logreg ~ season + holiday + workingday + weather + temp + atemp + humidity + windspeed + Ideal + hour + MONTHS + days + officehour + freehourworkday + weekend + ozone + light + dom, data = biketrain)</pre><pre># Predict</pre><pre>predict2 &lt;- predict(bkmodel1, biketest)<br></pre><pre>biketest$logreg = predict2<br>biketest$logcas = predict1</pre><pre>biketest$casual = exp(biketest$logcas)<br>biketest$registered = exp(biketest$logreg)</pre><pre>## Saving submission file as per requirement of competition</pre><pre>biketest$count=biketest$casual+biketest$registered<br>bike&lt;-data.frame(datetime=biketest$datetime,count=biketest$count)<br>write.csv(bike,file=&quot;bikeshare.csv&quot;,row.names=FALSE)</pre><p>This is Exploratory Data Analysis is for beginners and can be made more complex. These results can take you in the top 20 percentile of Kaggle competition, which is quite good for a beginner playground competition. Few more elements like ntree, importance, npreb, and proximity can be added to the random forest model to increase its efficiency.</p><p>Orignally published as a <a href="https://www.kaggle.com/rutvikdeshpande/bike-sharing-demand">Kaggle notebook</a>.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=32682a7a07e7" width="1" height="1" alt=""><hr><p><a href="https://medium.com/analytics-vidhya/predicting-bike-sharing-demand-using-randomforest-algorithm-32682a7a07e7">Predicting Bike Sharing Demand using RandomForest Algorithm</a> was originally published in <a href="https://medium.com/analytics-vidhya">Analytics Vidhya</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Site Analyser]]></title>
            <link>https://blog.devgenius.io/site-analyser-3a659227f776?source=rss-7970904ad169------2</link>
            <guid isPermaLink="false">https://medium.com/p/3a659227f776</guid>
            <category><![CDATA[entrepreneurship]]></category>
            <category><![CDATA[programming]]></category>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[architecture]]></category>
            <category><![CDATA[startup]]></category>
            <dc:creator><![CDATA[Rutvik Deshpande]]></dc:creator>
            <pubDate>Fri, 26 Jun 2020 09:34:52 GMT</pubDate>
            <atom:updated>2020-07-18T16:36:05.858Z</atom:updated>
            <content:encoded><![CDATA[<h4>Architectural Intelligence</h4><h4>Site Analysis Tool for Architects made using R Language</h4><blockquote>“I see my buildings as pieces of cities, and in my designs I try to make them into responsible and contributing citizens.” — César Pelli</blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*86F8G_NMameuzKNuGQThvQ.jpeg" /><figcaption>Photo by <a href="https://unsplash.com/@andrewtneel?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Andrew Neel</a> on <a href="https://unsplash.com/s/photos/maps?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Unsplash</a></figcaption></figure><p><strong>Site Analysis</strong> is a part of pre-designing process and is defined as a detailed study of geographical, climatic, historical and legal of piece of land where we are planning to design the space. It involves getting information about topographies, trees around, streets, connectivity to major locations, weather, climate and pollution, and then analysing this given data or information.</p><h4><strong>Site Analysis </strong>involves 3 important steps :</h4><ul><li><em>Research phase</em></li><li><em>Analysis phase</em></li><li><em>Synthesis phase</em></li></ul><p>Whenever we are done with selecting the site we must always analyse and check the locality and neighbourhood, for example green places in that specific area, vehicle population, which might be the reason for air and noise pollution, crime rate of the locality, proximity to nearest restaurants, cafes, schools, general stores, medical shops, vegetable markets, gyms, clinics and parks.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*J4IyQEMs_yc3reYEhJ1TOw.jpeg" /><figcaption>Photo by <a href="https://unsplash.com/@tbshg?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Tobias</a> on <a href="https://unsplash.com/s/photos/maps?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Unsplash</a></figcaption></figure><p>So I created a tool using Programming for Site Analysis. I programmed the tool in R language using one of its popular package <em>“leaflet”</em>. This tool gave me an advantage of locating a site offline, I can draw any shape on the map to show and define my site, it gives me the area, perimeter and longitude &amp; latitude of the centre of the site, I can highlight the site using different color and opacities, I can pin any location on the map, I can get pollution details and it gives me around 40 map provider options(like OpenStreetMap, ESRI, CartoDB, etc.) to choose from.</p><p>Using <strong>addMeasure()</strong> , the measuring tool is added. Primary and secondary units of area and length can be specified here. Also the color of the polygon before and after completion of the measurement can be specified.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/762/1*6Z_73hlB1yzyMbVXVKmIYw.png" /><figcaption>Map is ready !</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*deQvZrBESaDoZNsVwfW1tw.png" /><figcaption>ESRI provider is selected and satellite map is chosen to get the information about man-made and natural features</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*d_7tmRS6M1l0sEGLNx4Unw.png" /><figcaption>OpenTopoMap which gives the information about physical features</figcaption></figure><p>The tool gives me an option of 40 maps to choose from, like a topography map, relief features map, map with contour lines, watercolor map, dark map, grey map, bike density map, weather, and temperature map. These are maps from top map providers.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/682/1*QIXEmIRaNak9UXK02JwAWw.gif" /><figcaption>Creating a site using polygon drawing tool and then placing a location marker</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/600/1*28jaIaAG3-Cgbi_hzYEbxA.gif" /><figcaption>Changing color of the polygon and boundary line, specifying different opacity values of polygon, Customising the icon and its color</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/600/1*O-3qmfd49YCtrSILrKhYTA.gif" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/600/1*NFujUCDG9ZTM7d0n-Tm4tg.gif" /><figcaption>Measurement tool can be used for measuring the distance as well</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*qV30UJEPfdvfI3DRBWWpDw.png" /><figcaption>Perimeter and Area in meters as specified, with secondary unit as acres for area</figcaption></figure><p>This is a simple tool with many functions and easy too, though more complexities can be added to it. I created it because I thought instead of using a paid version of any site analysis software, I can make easy to use, simple and free tool.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=3a659227f776" width="1" height="1" alt=""><hr><p><a href="https://blog.devgenius.io/site-analyser-3a659227f776">Site Analyser</a> was originally published in <a href="https://blog.devgenius.io/">Dev Genius</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Beginning of My Data Science Journey]]></title>
            <link>https://blog.devgenius.io/beginning-of-my-data-science-journey-e887431ba834?source=rss-7970904ad169------2</link>
            <guid isPermaLink="false">https://medium.com/p/e887431ba834</guid>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[learning]]></category>
            <category><![CDATA[education]]></category>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[journey]]></category>
            <dc:creator><![CDATA[Rutvik Deshpande]]></dc:creator>
            <pubDate>Mon, 22 Jun 2020 17:33:41 GMT</pubDate>
            <atom:updated>2020-06-27T17:31:31.631Z</atom:updated>
            <content:encoded><![CDATA[<h4>A Learning Journey</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*K6v1lw7W34IYp6XtQoDaAg.jpeg" /><figcaption>Photo by <a href="https://unsplash.com/@impatrickt?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Patrick Tomasso</a> on <a href="https://unsplash.com/s/photos/journey?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Unsplash</a></figcaption></figure><p>I started developing interest in <em>Data Science</em> back in march 2020 and the journey of learning has began since then, but many people whom I meet and tell, that I am learning <em>Data Science</em> and am quite enthusiastic about it, they start to question me, “Then why have you chosen architecture as your major and why not computer science or any other major related to computers or software or programming ? Did your parents force you take architecture ?” Well the answer to these all questions is NO. Obviously there is scope in <em>Data Science</em> but I had no intentions of learning it just for the sake of trend or anything. Nobody forced me to learn either of the two subjects, Architecture or <em>Data Science</em>, both are my interests and I equally love studying both of them.</p><p>I was reading “Data-Driven Design and Construction”, A book by Randy Deutsch and one of it’s paragraphs motivated me to continue my learning, the paragraph was :</p><blockquote>Data is recognized by many in the architecture, engineering, construction, and operations (AECO) industry as the elephant in the room. Data and especially the catchall term big data is an important topic and, specific labels aside, is poised to remain so. To their credit, many design and AECO industry professionals already realize that data is the answer to their most perplexing professional and business problems.</blockquote><p>Also it is believed that amount of data available in AECO industry is the largest compared to other industries, but it is yet to be fully discovered, processed and used. I thought it is a great oppurtunity to begin <em>Data Science</em>.</p><p>I developed genuine interest when I got to know about the use cases of <em>Data Science</em>. When I say <em>Data Science</em>, it is the broader term which includes Machine Learning, Database Management, Statistics , Probability and of course Programming. While learning <em>Data Science</em>, I got to know that it can be divided into 6 simple parts :</p><ol><li><em>Importing the available data</em></li><li><em>Managing and Cleaning the data</em></li><li><em>Exploratory Analysis of data</em></li><li><em>Visualizing the data</em></li><li><em>Model Training and Evaluation</em></li><li><em>Deploying the Model</em></li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*vmC8lJ0rF8n4x9-DFXLlyQ.png" /></figure><h3>Learn</h3><p>Many people think <em>Data Science</em> is only about programming and only those can succeed who have exceptional programming skills, yes you do require basic programming knowledge either in Python or R (As these two Programming languages are widely used for <em>Data Science</em>), But <em>Data Science</em> is also about problem solving, logical reasoning, storytelling and mathematics.</p><p>So started learning the R programming language and did various online courses for it. I found pretty good courses for basic and intermediate programming at <a href="https://app.pluralsight.com/library/free">Pluralsight</a>, <a href="https://www.edx.org/">edx</a>, <a href="https://www.datacamp.com/">Datacamp</a> and <a href="https://www.udacity.com/">udacity</a>. To get into more depth of some of the topics, I referred some books by few of the renowned publishers <a href="https://www.packtpub.com/">Packt&gt;</a>,<a href="https://www.oreilly.com/">O’Reilly</a> and <a href="https://www.springer.com/in/search?query=DATA+SCIENCE&amp;submit=Submit">Springer</a>.</p><p>For Statistical Analysis and Machine Learning I had to brush my math skills — Statistics, Probability and Linear Algebra, though to understand you are just required to learn basic and few intermediate concepts of above mentioned. Lots of courses are available online but I just revised some important concepts.</p><p>Machine Learning is a major subset of <em>Data Science</em>. So after learning basic programming, Exploratory Data Analysis and Visualisation, I moved on to ML. I took various courses from renowned and popular online educational platforms like edx, <a href="https://app.pluralsight.com/library/free">Pluralsight</a>, <a href="https://www.datacamp.com/">Datacamp</a>, <a href="https://data-flair.training/">data-flair</a>. Luckily <a href="https://app.pluralsight.com/library/free">Pluralsight</a> had declared 7000+ of their courses free on their platform for the month of April and I took advantage of that. YouTube videos helped me a lot, I referred to these YouTube channels for more knowledge <a href="https://www.youtube.com/watch?v=4R8nWDh-wA0&amp;list=WL&amp;index=35&amp;t=2446s">Edureka</a>,<a href="https://www.youtube.com/channel/UCuWECsa_za4gm7B3TLgeV_A">Dr.Bharatendra Rai</a>(Great explanation of R concepts) and <a href="https://www.youtube.com/user/Simplilearn">Simplilearn</a>.</p><p>(I did not focus on earning a certificate as I just wanted to learn the concepts, so I used to audit the course)</p><h3><strong>Practice</strong></h3><p>Practice makes a man perfect is no more just a saying but you do improve yourself after practising. I started doing some beginner projects on my own, as there are lots of open datasets available, I love playing with them and trying out new things. <a href="https://app.pluralsight.com/library/free">Pluralsight</a> offers a platform called Skills IQ where you can test your skills.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*a5OfFGPH6WQJNafjBkFOJA.png" /><figcaption><a href="https://app.pluralsight.com/library/free">Pluralsight</a></figcaption></figure><p>It gives you a great insight about what is yet to be covered, what are you good at (I still have much to learn and cover about as you can see it in my assessment). I took the <a href="https://www.datacamp.com/">Datacamp</a> assessments too and they also seemed to be pretty good.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*3luwugcEdwy6VtAFOg7elw.png" /></figure><h3>Explore</h3><p>This is one of the optional processes involved in the data science journey, though it offers great benefits. When I started competitive coding at <a href="https://www.kaggle.com/">Kaggle</a>, initially I performed poorly like getting into top 98 percentile of one of its competition and I got sad. But soon after that I started viewing other notebooks in <a href="https://www.kaggle.com/">Kaggle</a>, and I must say I learned a lot from it. There were few ML algorithms which I did not learn in any of the courses, like Gradient Boosting or Pruning or say few Meta-Algorithms. <a href="https://stackoverflow.com/">Stack Overflow</a> and <a href="https://rpubs.com/">RPubs</a> helped me a lot in solving my doubts. I follow a few podcasts like <a href="https://www.oreilly.com/radar/topics/oreilly-data-show-podcast/">O’Reilly</a> and <a href="http://www.ocdevel.com/mlg">OCDevel</a>.</p><p>Reading has helped me a lot in this beautiful journey and it’s greatest source was <a href="https://medium.com/">Medium</a>, I joined medium back in April and used to read 1–2 articles a week initially. After that I started saving stories for future references, Now I even save stories of the concepts which I have not yet learned. Reading about others projects gives you great inspiration. Reading stories on <a href="https://medium.com/">Medium</a> motivated me to write my <a href="https://medium.com/swlh/analyzing-and-predicting-manchester-uniteds-performance-a-data-science-perspective-971afeb9c1cf">first article on sports and data science</a>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*KVVj726C6hq0IeMfLUD1Yw.png" /></figure><h3>Compete</h3><p>Competition is always a good thing, it forces us to do our best. Competitive programming is crucial learning path of data science. I started my competitive programming at <a href="https://www.kaggle.com/">Kaggle</a>. And like anyone else, I failed a couple of times(even got a percentile of 98). For my first playground competition, <a href="https://www.kaggle.com/c/titanic">Titanic: Machine Learning from Disaster</a>, I submitted 13 entries and with the 13th one I got ranked in top 9 percentile. I also tried a few other competitive programming platforms like HackerEarth and AnalyticsVidhya but I am not as consistent there as I am on <a href="https://www.kaggle.com/">Kaggle</a>. Competitive programming taught me many things but the the key take away is “A same problem has many different approaches.”</p><blockquote>“TELL ME AND I FORGET. TEACH ME AND I REMEMBER. INVOLVE ME AND I LEARN”–BENJAMIN FRANKLIN</blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*J4SZQ1qnDWXqv6DB74IlYA.png" /></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e887431ba834" width="1" height="1" alt=""><hr><p><a href="https://blog.devgenius.io/beginning-of-my-data-science-journey-e887431ba834">Beginning of My Data Science Journey</a> was originally published in <a href="https://blog.devgenius.io/">Dev Genius</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Creating a Unique and Beautiful Presentation using R Programming Language]]></title>
            <link>https://medium.com/swlh/creating-a-unique-and-beautiful-presentation-using-r-programming-language-3dd9e3cc10fe?source=rss-7970904ad169------2</link>
            <guid isPermaLink="false">https://medium.com/p/3dd9e3cc10fe</guid>
            <category><![CDATA[programming]]></category>
            <category><![CDATA[design]]></category>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[learning]]></category>
            <category><![CDATA[presentations]]></category>
            <dc:creator><![CDATA[Rutvik Deshpande]]></dc:creator>
            <pubDate>Mon, 22 Jun 2020 08:41:37 GMT</pubDate>
            <atom:updated>2020-06-22T14:12:57.779Z</atom:updated>
            <content:encoded><![CDATA[<blockquote><em>“A great presentation gives smart ideas an advantage”</em></blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Q3Nb34zq37PvtL1Fk_eB4A.png" /></figure><p>I was interning with an AI/ML startup — <a href="http://rises.io/">Rises Analytics Solutions Pvt Ltd</a>, in my hometown when the CEO of the company got invited by a college for a presentation of the topic, <strong>AI vs COVID-19</strong>, for their students. Since I had just joined the firm for internship I was assigned to make that presentation, which was a great experience in itself and got me to learn a lot.</p><p><strong>So first of all let us know what exactly is a presentation.</strong></p><p>A presentation is, typically a demonstration, introduction, lecture meant to inform, persuade, inspire or to present a new idea or product it is basically visual storytelling. So while narrating a story you have a certain sequence of events. Not necessarily but a story has a few elements without which it seems incomplete, like a Point of view, Perspective, Morale and many more things. These are all relevant in a presentation too, you need to slowly build up the story and main topic of the presentation.</p><p><strong>What have I done ?</strong></p><p>What have I done, why am I sharing my experience about a normal presentation. Well this was not a normal presentation, I made the whole presentation in R programming Language using one of it’s javascript libraries — reveal.js, an open source HTML presentation framework. It’s a tool that enables anyone with a web browser to create fully-featured and beautiful presentations. Presentations made with reveal js are built on open web technologies.</p><p>To create a reveal.js presentation from R Markdown you specify the revealjs_presentation output format in the front-matter of your document.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/532/1*j6fEkJIHkXOqb8Q4-mtLZQ.png" /><figcaption>For more details and information, check out <a href="https://garrettgman.github.io/rmarkdown/revealjs_presentation_format.html">reveal.js</a></figcaption></figure><p>You can create a slide show broken up into sections by using the # and ## heading tags (you can also create a new slide without a header using a horizontal rule (----). Here is the starting code :</p><pre>---<br>title: &quot;AI against COVID-19&quot;<br>author: &quot;Name of the Author&quot;<br>date: 24-04-2020<br>output:<br>  revealjs::revealjs_presentation:<br>    theme: night<br>    highlight: espresso<br>    center: true<br>    transition: concave<br>---</pre><p><strong>Who was the target audience here ?</strong></p><p>It was a presentation for engineering students interested in AI,ML and NLP . Mainly students with majors in Computer Science and Information Technology.</p><p><strong>Why did I choose R and not any other platforms like Microsoft Powerpoint Presentation ?</strong></p><p>I am interested in Data Science and for that I am learning the R Language, Since the audience were engineering college students and interested in programming, I was asked to use an open source software for making the presentation as it would make a lasting impression about the company and students would love a different approach to make a presentation.</p><p><strong>How did I make the presentation, which was appreciated by many participating students ?</strong></p><p>Coming back to the storyline of presentation topic — COVID-19, sequence of my story was</p><ol><li><em>Greetings and welcoming everybody. Tried to lighten the pressure on students, due to effects of the virus and pandemic, using a cartoon.</em></li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*tJihaDBnEnDv3Cqp" /></figure><p>2. <em>Information about the virus.</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*dWnHGWLaDe42duzc" /></figure><p>3. <em>Global effects.</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*tEbRQ04swbfRHLlu" /></figure><p>4. <em>Effects of virus on India.</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*45wd0sjvsyj446hs" /></figure><p>5. <em>Few slides of live graphs, of which some were made by me in R using the most popular ggplot library.</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/600/1*AnIJzZY1z3q2KSA4w0-Z6g.gif" /></figure><p>6. <em>How is AL being used to battle COVID-19.</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*YFNL9r-cwq-l1khX" /></figure><p>7. <em>Use Cases of ML, NLP in fighting COVID-19.</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*rHQ0i3x6O8-yIHq5" /></figure><p>8. <em>Top Global tech companies deploying AI.</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*7Ewg7lpi6VOj0Wz3" /></figure><p>9.<em>Top Indian tech companies working against COVID-19 using AI</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*o0j8zHCq1y299X9l" /></figure><p>10. <em>What is the Presenter’s company doing (Introduction of the company).</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*BKnvc-eBUq-Adu0k" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/600/1*wO2OminfmcIZiabnxgWCDw.gif" /><figcaption>Information Courtesy : <a href="http://rises.io/">Rises Analytics Solutions</a></figcaption></figure><p><em>11. Company’s contribution in tackling the Pandemic Situation and the virus.</em></p><p><em>12. How can students contribute in this pandemic situation</em>.</p><p>It is alright to tell what tech giants are doing and what is the situation all over the world, but students will have a doubt, how can they contribute. So I provided certain competition information.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*4p3FHEHhLrnLr0vk" /></figure><p><em>13. Information about the presenter.</em></p><p><em>14. Ending the presentation on a note asking them to stay safe and healthy amid COVID-19 situation.</em></p><p>When you see the basic plot of the presentation, the presentation starts with broader and general information about the topic as it is important that whoever is attending the presentation must have an basic idea about the topic. This is presented in two segments — Global effects and Local effects. Then we slowly move on to the agenda of the presentation, followed by examples again Global and then Local. Further there is information about the company and then their contribution. Ending with “What can be your role”.</p><p><strong>What issues I faced while making this presentation</strong></p><p>R is certainly not a simple programming language to use, also it was my first time making a presentation in R, I did face certain issues, which were as follows :</p><ol><li>Centre = TRUE is default in reveal js, this was one of the major issues I faced while making the presentation. Initially I was not able to make columns in a slide for presenting the content, like presenting the pictures and written information together.</li><li>Secondly there was a problem, lag during playing the presentation, whenever a specific slide was played for a longer time, there was a lag of 30–40 seconds. I could not set an automatic time limit for playing the slides as it depends on the presenter, how much time he or she will speak about that slide.</li></ol><blockquote>“<em>The success of your presentation will be judged not by the knowledge you send but by what the listener receives.”</em></blockquote><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=3dd9e3cc10fe" width="1" height="1" alt=""><hr><p><a href="https://medium.com/swlh/creating-a-unique-and-beautiful-presentation-using-r-programming-language-3dd9e3cc10fe">Creating a Unique and Beautiful Presentation using R Programming Language</a> was originally published in <a href="https://medium.com/swlh">The Startup</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>