<?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[ZebraX - Medium]]></title>
        <description><![CDATA[We are #ZebraX, a digital transformation company to provide integrated solutions from data collection, processing, and producing actionable insight using advanced analytics and machine learning model specifically made for industrial use cases. - Medium]]></description>
        <link>https://medium.com/zebrax?source=rss----42d4c8441b9a---4</link>
        <image>
            <url>https://cdn-images-1.medium.com/proxy/1*TGH72Nnw24QL3iV9IOm4VA.png</url>
            <title>ZebraX - Medium</title>
            <link>https://medium.com/zebrax?source=rss----42d4c8441b9a---4</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Wed, 18 Nov 2026 07:33:08 GMT</lastBuildDate>
        <atom:link href="https://medium.com/feed/zebrax" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[What is a Beautiful Code?]]></title>
            <link>https://medium.com/zebrax/what-is-a-beautiful-code-b242c2ed219b?source=rss----42d4c8441b9a---4</link>
            <guid isPermaLink="false">https://medium.com/p/b242c2ed219b</guid>
            <category><![CDATA[coding]]></category>
            <category><![CDATA[programming]]></category>
            <category><![CDATA[engineering]]></category>
            <category><![CDATA[software-engineering]]></category>
            <category><![CDATA[code]]></category>
            <dc:creator><![CDATA[ZebraX Engineering]]></dc:creator>
            <pubDate>Thu, 07 Apr 2022 06:59:38 GMT</pubDate>
            <atom:updated>2022-04-07T06:59:38.058Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*GgkeLUc7QseAF_d4UY9Xnw.png" /></figure><p>What do you think is a beautiful code? Is the code neatly organized? Is the code colorful and easy to read? There will be many opinions about this.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1000/0*R2TN51ouqfipVoG_" /></figure><p>The traditional idea of beauty is based on the physical appearance of things. To be beautiful is to be visually appealing. However, when it comes to code, the word “Beautiful” is more complex than you would think. Beautiful code has to be easily comprehensible and also syntactically correct.</p><p>So basically, beautiful code is a code that can make everyone understand what this code is for, even when the code is read again after years and read by other people. This is something that is very helpful for the smooth distribution of this code understanding.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/914/0*5idfrEnwIOycMFQ3" /></figure><p>I’ve had a really good experience seeing code samples like this. This happened the first time I did my internship at ZebraX. I’m able to understand the code faster than I expected. I can understand every logic quickly with just every component in it.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/552/0*o3QAp5Qap2V2uEss" /></figure><p>There are a lot of things that contribute to what makes a code beautiful.<br>Here are some criteria for beautiful code</p><p><strong>Simple<br></strong>Every code needs a reason to exist, otherwise, it has to be refactored or deleted and every piece of code must have an unambiguous purpose. In other words, simply reducing the code while maintaining functionality</p><p><strong>Simply written<br></strong>Simplifying complexity makes the code easier to understand.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/572/0*envGVQ3FtPooSV5k" /><figcaption>Example for <strong>Simply Written</strong></figcaption></figure><p><strong>Self-documenting<br></strong>Beautiful code needs relatively few comments because it’s so clear that you can read it like a book</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/475/0*JlUoFj91kMPOKCrD" /><figcaption>Example for <strong>Documenting</strong></figcaption></figure><p><strong>Modular</strong><br>Modularization allows you to reuse parts of your code, to achieve it you need to consolidate a few things which are:</p><ol><li>DRY (Don’t Repeat Yourself)<br>A principle aimed at reducing repetition of software patterns, it’s mandatory to generalize and avoid repetition to avoid redundancy.</li><li>YAGNI (You Aren’t Gonna Need It)<br>Minimize the number of entities like functions, classes, modules, etc. A programmer should not add functionality until deemed necessary. If you have a lot of nearly identical functions, the best practice is to refactor or combine them.</li><li>SDOT (Should Do One Thing) <br>A function will be difficult to generalize and reuse if it’s doing multiple things</li><li>Try to use fewer arguments per function<br>In many cases, it’s more effective to use fewer arguments. The reason is not only it’s easier to read and understand but also makes the code more efficient to work with.</li></ol><p><strong>Error Handling<br></strong>Handles errors consistently, clearly, correctly and in a way that aids debugging</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/350/0*FnYM4L_4IrVef6al" /><figcaption>Example for <strong>Error Handling</strong></figcaption></figure><h3>Does beautiful code impact your quality as a developer?</h3><p>Your quality as a developer is determined by how robust, effective, efficient, and easily comprehensible your code is. Writing beautiful code would also reduce or even avoid technical debt. Beautiful code will be much easier to maintain and read which can affect your project in the long run.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/900/0*gzjXLfOTJmUt1bTP" /></figure><h3>Conclusion</h3><p>Beautiful code is one of the important things to implement because it’s very helpful to deliver understanding between developers in a project. It makes the code more efficient and reusable. Maybe it looks like a simple thing but in reality, beautiful code is very useful to practice.</p><p><strong>Author:</strong></p><ul><li>Aryasatya Bhagaskara (Software Engineer)</li><li>Randa Alfarisyi (Software Engineer)</li></ul><p><strong>Source:</strong><br><a href="https://medium.com/@fahdmekky/writing-clean-and-modular-code-eb1eb8bba87f">https://medium.com/@fahdmekky/writing-clean-and-modular-code-eb1eb8bba87f</a><br><a href="https://pavlepesic.medium.com/beautiful-code-principles-39420873eff8">Beautiful Code Principles | by Pavle Pesic | Medium</a><br><a href="https://www.tabnine.com/blog/what-makes-code-beautiful-and-why-it-matters/">What makes code beautiful and why it matters (tabnine.com)</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=b242c2ed219b" width="1" height="1" alt=""><hr><p><a href="https://medium.com/zebrax/what-is-a-beautiful-code-b242c2ed219b">What is a Beautiful Code?</a> was originally published in <a href="https://medium.com/zebrax">ZebraX</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Leveraging Geospatial Framework in Big Data Analytics]]></title>
            <link>https://medium.com/zebrax/leveraging-geospatial-framework-in-big-data-analytics-3fe77f15ff9c?source=rss----42d4c8441b9a---4</link>
            <guid isPermaLink="false">https://medium.com/p/3fe77f15ff9c</guid>
            <category><![CDATA[data-management]]></category>
            <category><![CDATA[geographic-information]]></category>
            <category><![CDATA[geospatial-analytics]]></category>
            <category><![CDATA[big-data-analytics]]></category>
            <category><![CDATA[data-analytics]]></category>
            <dc:creator><![CDATA[Gilang Samudrakuswana]]></dc:creator>
            <pubDate>Fri, 11 Mar 2022 05:03:31 GMT</pubDate>
            <atom:updated>2022-03-11T09:23:11.289Z</atom:updated>
            <content:encoded><![CDATA[<h4>One of the efforts that can support Sustainable Development Goals through data management</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Q9IZSMfZRqYQIbr4JJENmg.jpeg" /></figure><blockquote>There are many ways that industries can do to support Sustainable Development Goals (SDGs) set by the United Nations (UN), in order to achieve the 2030 agenda. One of the efforts that can be done is by utilizing Geographic Information Systems (GIS) in Big Data Analytics. GIS can help to support Sustainable Development Goals through data management, analysis and understanding, increased collaboration, sustainable design and planning, and sound decision making.</blockquote><h3>The need for a big data environment in geospatial analytics</h3><p>Today, the volume of geospatial data increases drastically and faster than ever. As we all already know, performing geospatial analysis may call for special tools to analyze spatial data to generate insights. Many users use classic geographic information systems (GIS), specialized software systems for storing, managing, analyzing, and visualizing geospatial data. The classic GIS is not designed for this huge data, it is primarily designed for data that fit in a single machine.</p><p>One of the main reasons why today, we have a huge amount of geospatial data, and still generating new data every day, is because of the new trend and advanced technology in mobile devices and geolocation sensors. We can find them everywhere from our devices to industrial equipment and vehicles. All of them are generating geospatial data (Spatio-temporal data).</p><p>This huge amount of geospatial data is a great potential of valuable information that can be used for any kind of application: industry, environment, retail, and many more. Below are examples of geospatial applications that have already been implemented widely:</p><ul><li>Climate modeling and analysis</li><li>Urban planning</li><li>Consumer insights</li><li>Route optimization</li><li>Insurance and Fraud analysis</li></ul><p>The massive development of distributed, parallel computing with commodity hardware or cloud enables access to more computing power that is reliable, scalable, easy to launch, and cheaper. The big data platforms, with all of their potential, now pave the way for geospatial big data analytics.</p><p><strong>Meet Apache Sedona</strong></p><p>The Apache Software Foundation incubator runs a project to extend the core engine of Apache Spark and SparkSQL to support spatial data types, indexes, and geometrical operations at scale. This new platform is called Apache Sedona (formerly GeoSpark). Apache Sedona is created and architected by Dr. Mohamed Sarwat, an expert on database systems and spatial analytics who is an assistant professor of computer science at Arizona State University.</p><p>Apache Sedona is a cluster computing system for processing large-scale spatial data. Sedona provides a set of Spatial Resilient Distributed Datasets (SRDD) and Dataframes that can efficiently load, process, and analyze large-scale spatial data across a cluster of machines. Sedona provides Scala, SQL, Java, and Python APIs to program analytics pipelines on-top of geospatial data. Apache Sedona is a general-purpose geospatial data processing engine. Using Sedona, we can scale our geospatial works across many machines and we can run it everywhere in the cloud such as Amazon EMR, GCP data proc, Databricks, or on-premises Hadoop cluster.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*lSvgza5FVqhVF0lq" /><figcaption>System Architecture of Apache Sedona</figcaption></figure><p>A major advantage of Apache Sedona is its ability to load data into memory and physically split in-memory copies to a lot of equally sized partitions across work nodes of a cluster, while still preserving spatial proximity of the data. This is done by grouping spatial objects based on their spatial proximity. This leads to fast query speeds.</p><p><strong>Supported Data Sources</strong></p><p>Apache Sedona provides the possibility of loading the data from various data sources such as CSV, TSV, WKT, WKB, GeoJSON, Shapefile, PostGIS, spatial parquet, and NetCDF / HDF format. Spatial RDDs can accommodate seven types of spatial data including Point, Multi-Point, Polygon, Multi-Polygon, Line String, Multi-Line String, GeometryCollection, and Circle.</p><h3>Application in Shipping Industry</h3><p>For maritime operations, geolocation plays an important role. It is important in marine navigation for the ship’s officer to know the vessel’s position while in the open sea and also in congested harbors and waterways. While at sea, accurate position, speed, and heading are needed to ensure the vessel reaches its destination in the safest, most economical, and timely fashion that conditions will permit.</p><p>The implementation of tracking devices such as GPS, allows organizations to collect time and place based on geo-references in any event. By enabling organizations to link their asset location, it opens the opportunity to asset monitoring, generating insights, and building analytics models using the available data. The geospatial analysis uses the data to build maps, graphs, and statistics to make complex data more understandable.</p><p>In this article, we will look into geofencing application and vessel activity prediction use cases for maritime operations. The objective of geofencing is to give an alert when a vessel exits a virtual boundary set up around a geographical location. The second use case is aimed to help ship officers in making timesheets by predicting the vessel’s activity every 15 minutes, so it generates the timesheet automatically and the ship officer can focus on more important tasks.</p><p><strong>Data visualization</strong></p><p>The most interesting part of the geospatial analysis is that it is highly visual. So, let’s make some visualizations to get a better understanding of the data. The picture below is the route that will be assigned to the asset. In this example we show you 3 routes i.e. “Kalimantan 1 — Jawa 1”, “Sumatra 1 — Jawa 2”, and “Sumatra 1 — Jawa 3”. Normally, all vessels should follow these routes unless there are some conditions that do not allow them to.</p><p>*you can play with the map by zooming out/in to see a more detailed location of the data</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fdatapane.com%2Fu%2Fuser616%2Freports%2F0AEjpp3%2Fvessel-tracking-dist%2Fembed%2F&amp;display_name=Datapane&amp;url=https%3A%2F%2Fdatapane.com%2Fu%2Fuser616%2Freports%2F0AEjpp3%2Fvessel-tracking-dist%2F&amp;image=https%3A%2F%2Fstorage.googleapis.com%2Fdatapane-files-prod%2Fdp%2Fthumbnails%2Fe6a0505f-dcf2-4fea-9c21-7fc068e9184a.png&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=datapane" width="960" height="700" frameborder="0" scrolling="no"><a href="https://medium.com/media/c395995ac998fc37628e2ab8c16497ae/href">https://medium.com/media/c395995ac998fc37628e2ab8c16497ae/href</a></iframe><p>From the data points distribution, it is apparent that the port area is denser than any other area. This is due to at a port, vessels are doing loading/unloading activity which takes some time to finish. So basically, we will have more data points around the port area with near-zero movement.</p><p>Another insight that we found from the map is that we have outliers around the South Sumatra area. Our vessels are not assigned to that area. It possibly has resulted from a glitch in our devices.</p><h3>Geofencing</h3><p><strong>Deviation Measurement</strong></p><p>To build an alert system, we need to specify the trigger to activate the system. In our case, the deviation distance from the original route should be a good metric, but sometimes deviation from the original route is inevitable for special cases like extreme weather conditions.</p><p>The picture below shows you the original routes overlayed with the vessel position and its color-coded deviation distance information.</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fdatapane.com%2Fu%2Fuser616%2Freports%2F43gjEYA%2Findonesia-vessel-tracking%2Fembed%2F&amp;display_name=Datapane&amp;url=https%3A%2F%2Fdatapane.com%2Fu%2Fuser616%2Freports%2F43gjEYA%2Findonesia-vessel-tracking%2F&amp;image=https%3A%2F%2Fstorage.googleapis.com%2Fdatapane-files-prod%2Fdp%2Fthumbnails%2Faf4b8b53-3b3b-4e61-bba6-5b5897dda190.png&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=datapane" width="960" height="700" frameborder="0" scrolling="no"><a href="https://medium.com/media/7665388d8de3cf16b55f4f04b2947ae4/href">https://medium.com/media/7665388d8de3cf16b55f4f04b2947ae4/href</a></iframe><p>From the 3 routes, the most interesting finding is in the “Kalimantan 1 — Jawa 1” route. The map shows high and frequent deviations from our vessels. From the pattern, the deviation is quite consistent and it creates a pattern resembling a new different route. This is the statistical value of the deviation.</p><ul><li>Route name: “Kalimantan 1 — Jawa 1”<br>Average deviation (meter): 24760.914<br>Standard deviation (meter): 24840.5</li></ul><p>Given the result, there is the possibility that we have a new route for the vessel that travels from “Kalimantan 1” to “Jawa 1”. This is a good finding. We can bring this information to the shipping company and tell them that their current system is possibly not updated for the new route.</p><p><strong>Incorporating Time for Deviation measurement</strong></p><p>As stated earlier, sometimes deviation is inevitable for a special case like weather conditions. If a captain decides to deviate from the original route, we may allow such conditions to happen as long as the ship returns once the weather is better. One simple idea to accommodate this is by incorporating the time-window variables into the alert system. We can dynamically set the time window to 30 minutes, 1 hour, or even more, as long as each stakeholder agrees upon the number. Within this time window, the alarm will not be set on, but after that, the alarm will notify as usual.</p><h3>Ship Activity Prediction</h3><p>The objective is clear and the solution seems simple. If we have the vessel’s location then we possibly know what activity is being carried out, since basically activities are bound to the geolocation references. It is uncommon to do loading/unloading activity in the middle of the ocean, we do this kind of activity at the port.</p><p>We have 36 different activities, and here is the scatter plot of vessel locations and their corresponding activities. Normally we expected some groups of activities with specific locations will emerge. But, no. It shows no pattern at all. It looks like a complete mess, and impossible to derive rules from that (using clustering technique nor human interpretation from the visualization).</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/580/0*btVDMZDbrD4xSJqO" /></figure><p><strong>Modeling approach</strong></p><p>The case is a bit more complex than we thought, so we build a predictive model instead of deriving rules. The supervised technique is suitable here since we have all the labels. Below are the general steps that we followed in the modeling process.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*NAs18FIbmZEI8s98" /></figure><p>We used 28 features including time-series features to capture the vessel’s movement in the last 2 hours. Highlighted with a yellow box below.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/146/0*xBhe8lw8Ogc2qFXc" /></figure><p>Having all the requirements, we built a predictive model using ZebraX advanced analytics platform (called ZX AnalytiX). No need to hard code the pipeline, just click and specify the algorithm, hyperparameters, feature engineering, and filtering process that we want to take.</p><p>Here is a screenshot of the experiment report of our modeling process using the Random Forest algorithm. The f1 score of our model is quite good for the first trial, 0.7. From the graph, it seems there is no indication of overfitting.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*PXrwTGIkaShJzmS9" /></figure><h3>Summary</h3><p>In the context of the geofencing system, this article shows you that we can complement the existing tracking system with additional logic that will give us an alert when our vessel is exiting our virtual boundary and also context-aware in which external information is rigorously analyzed before alerting the user, thus making the alarm smart.</p><p>The other potential that we show in this article is an automated report system that will help the ship officers to make a more consistent and accurate timesheet. All the time-consuming tasks in creating a report will be handled by a program, and they can focus on more important tasks.</p><p>👋 Thanks for reading.</p><p>Salam…</p><h3>Reference</h3><p><a href="https://sedona.apache.org/">https://sedona.apache.org/</a></p><p><a href="https://www.gislounge.com/geospatial-data-and-sustainable-development-goals/">https://www.gislounge.com/geospatial-data-and-sustainable-development-goals/</a></p><h3>Author</h3><ul><li>Gilang Samudra K. (Sr. Data Scientist)</li><li>Puteri Aulia Indrasti (Data Engineer)</li><li>Daniel Beltsazar M. (Data Scientist Intern)</li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=3fe77f15ff9c" width="1" height="1" alt=""><hr><p><a href="https://medium.com/zebrax/leveraging-geospatial-framework-in-big-data-analytics-3fe77f15ff9c">Leveraging Geospatial Framework in Big Data Analytics</a> was originally published in <a href="https://medium.com/zebrax">ZebraX</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Predictive Maintenance, Cut Costs & Improve Enterprise Operational Efficiency]]></title>
            <link>https://medium.com/zebrax/predictive-maintenance-cut-cost-improve-enterprise-operational-efficiency-c06536c76b41?source=rss----42d4c8441b9a---4</link>
            <guid isPermaLink="false">https://medium.com/p/c06536c76b41</guid>
            <category><![CDATA[free-trial]]></category>
            <category><![CDATA[digitaltrasformation]]></category>
            <category><![CDATA[operational-efficiency]]></category>
            <category><![CDATA[predictive-maintenance]]></category>
            <category><![CDATA[advanced-analytics]]></category>
            <dc:creator><![CDATA[ZebraX Team]]></dc:creator>
            <pubDate>Thu, 27 Jan 2022 05:44:17 GMT</pubDate>
            <atom:updated>2022-02-04T09:03:53.331Z</atom:updated>
            <content:encoded><![CDATA[<h4>Get our FREE Trial offer now!</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*cJNWLwjMKtyUHB_BQicsuw.jpeg" /></figure><blockquote>Climate change is one of the global issues today. Indika Energy, as a diversified energy company in Indonesia, also took a role to support the achievement of sustainable development goals (SDGs), targeting the achievement of carbon neutrality by 2050. Various efforts can be made to achieve these goals, one of which is by implementing the right technology so that the business can run more effectively and efficiently.</blockquote><h4>Technology, used in supporting businesses, has a role to play in achieving those sustainability goals. Because in fact, an efficient business is seen from the fewer carbon footprints it produces.</h4><p>ZebraX, a subsidiary of Indika Energy, also supports Indika Energy’s program through its digital solutions to help businesses be more productive and ready to compete in today’s digital transformation era. ZebraX’s flagship solutions include Advanced Analytics and IoT.</p><p>Several studies were conducted by industry to reduce carbon footprint, including by using synergies from Artificial Intelligence and Internet of Things (IoT) applications, namely by using analytics solutions-predictive maintenance use case.</p><p>The fundamental principle used by predictive maintenance is the ability to detect problems before events, to prevent the occurrence of unwanted things. However, industry players must also pay attention to the importance of maintaining consistency in the frequency of maintenance.</p><p>This exact frequency of maintenance is very important because it can lead to costs. The frequency that too often can cause maintenance costs to increase, when not necessarily needed. While maintenance that is rarely done can lead to the cost of loss, damage can even cause loss of the company.<br>Currently what needs to be known by the industry is how to determine when the right time to do maintenance. At this stage, the role of the synergy of Artificial Intelligence and the Internet of Things, through predictive maintenance can be the key to answering the problems faced by the industry.</p><h4>Benefits of Predictive Maintenance</h4><p>One of the main benefits that can be obtained if the industry uses predictive maintenance can reduce operational costs or maintenance costs. According to a survey conducted by IBM, predictive maintenance utilization can cut up to 40% of the total cost of maintenance.</p><p>In addition, predictive maintenance can also improve the safety of workers, both workers who perform maintenance and workers who operate maintained machinery. Predictive maintenance based on AI can help in empowering human resources more efficiently.</p><p>Other benefits of predictive maintenance can also address downtime and maintain uptime.</p><h4>The Steps of Predictive Maintenance</h4><p>In applying predictive maintenance, several stages must be done, namely data collection/data procurement, data preparation, feature engineering, training model, prediction, and deployment.</p><p>Data is the main element in applying AI. The data will later be implemented into the AI model to then be deployed anywhere or can be said to be flexible. So that later this AI model will be easily integrated with existing systems.<br>In addition, it is also necessary to detect hidden patterns. By using AI technology, detection can be done automatically and we can gain more insight. Unlike if done manually by humans, it will be difficult to detect data with large and very complex amounts.</p><p>And that needs to be underlined, this predictive maintenance process can not be done only once, but must be done continuously because of course we will continue to update the existing AI model, adjusted to the real case that occurs in the field so that the resulting insights will be better.</p><blockquote>The application of the right solutions and technologies can also be useful in achieving environmental sustainability to support Indika Energy’s commitment to creating a carbon-neutral in 2050.</blockquote><h3>30 Days Free Trial of ZebraX Advanced Analytics</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Lv0WpUos7peAlh9k0XDJ9Q.jpeg" /><figcaption>Free Trial DMAA</figcaption></figure><p>ZebraX, a digital transformation company, is ready to help improve the company’s operational efficiency by using advanced analytics — predictive maintenance solutions.</p><p>This solution can be implemented in various industrial sectors, including helping the manufacturing sector to predict when the maintenance time of machinery, know the results of production and inventory. In the retail sector, these solutions can read customer behavior.</p><p>In the health sector, the solution can support the diagnosis of doctors in hospitals. In the telecommunications and media sectors, the solution can predict customer experience and loyalty.</p><p>In the banking and financial sectors, the solution can predict customer behavior and in meeting regulations/compliance.</p><h4><strong>And right now we’re also offering free trials for companies looking to start their digital transformation journey using Advanced Analytics and also our IoT Platform solutions. For further information about this FREE Trial program, please contact us at inquiry@zebrax.id.</strong></h4><p>Source:</p><p><a href="https://www.ibm.com/services/technology-support/multivendor-it/predictive-maintenance">https://www.ibm.com/services/technology-support/multivendor-it/predictive-maintenance</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c06536c76b41" width="1" height="1" alt=""><hr><p><a href="https://medium.com/zebrax/predictive-maintenance-cut-cost-improve-enterprise-operational-efficiency-c06536c76b41">Predictive Maintenance, Cut Costs &amp; Improve Enterprise Operational Efficiency</a> was originally published in <a href="https://medium.com/zebrax">ZebraX</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Utilizing The Power of Typescript and GraphQL Codegen]]></title>
            <link>https://medium.com/zebrax/utilizing-the-power-of-typescript-and-graphql-codegen-bfb3924039ab?source=rss----42d4c8441b9a---4</link>
            <guid isPermaLink="false">https://medium.com/p/bfb3924039ab</guid>
            <category><![CDATA[graphql]]></category>
            <category><![CDATA[codegen]]></category>
            <category><![CDATA[software-development]]></category>
            <category><![CDATA[typescript]]></category>
            <category><![CDATA[software-engineering]]></category>
            <dc:creator><![CDATA[Fajar Septiawan]]></dc:creator>
            <pubDate>Fri, 26 Nov 2021 09:47:27 GMT</pubDate>
            <atom:updated>2021-12-01T07:23:57.261Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*2nPwUuAu8_qlOIxILMU2yQ.png" /></figure><p>In this article, we’ll try to tell our journey on using GraphQL and GraphQL Code Generator with Typescript on the Frontend Development team.</p><p><strong>The Idea of Using GraphQL Codegen</strong></p><p>Starting from the first time, we were still using Javascript in our project, then we moved to use Typescript since it supports static typing and is easier to read and understand. Also, it’s useful because this way you can detect potential bugs before they happen in runtime. We want to take advantage of static typing to convert GraphQL schemas into Typescript types and to allow our GraphQL Operations (<strong>query</strong>/ <strong>mutation</strong>/ <strong>subscription</strong>, and <strong>fragment</strong>) to select specific types from them using GraphQL Code Generator.</p><p><strong>Basic Fetching Data</strong></p><p>For example, let’s say we want to integrate an API that retrieves User List and use it on our React project using Typescript and Apollo Client by implementing types in each schema manually.</p><p>Let’s assume that we have a simple GraphQL query:</p><pre>// GET_USER_QUERY<br>query users {<br>  users {<br>   data {<br>    id<br>    name<br>    username<br>    email<br>  }<br> }<br>}</pre><p>Add the types of our <strong>query</strong></p><pre><em>// Types<br>export</em> interface UserData {<br> id: string<br> name: string<br> username: string<br> email: string<br>}</pre><pre><em>export</em> interface Users {<br> data: UserData[]<br>}</pre><pre><em>export</em> interface RootResponse {<br> users: Users<br>}</pre><p>And then we need to call the <strong>query </strong>in our client-side application:</p><pre><em>// React component<br>import</em> { useQuery } <em>from</em> &#39;@apollo/client&#39;<br><em>import</em> GET_USER_QUERY <em>from</em> &#39;~/queries/users&#39;<br>const User = () =&gt; {</pre><pre>const { data } = useQuery&lt;RootResponse&gt;(GET_USER_QUERY)<br> <em>// `data` is fully typed based on your GraphQL query<br>return</em> (<br>  */ Render */<br> )<br>}</pre><p>Everything works perfectly and all seems good and ready to go. Seems pretty simple right? Then how about this query:</p><pre>query usersWithPostsComments($options: PageQueryOptions){<br> users(options: $options){<br>  data {<br>   id<br>   username<br>   email<br>   name<br>   posts {<br>    data {<br>      title<br>      body<br>      comments {<br>       data {<br>        name<br>        email<br>        body<br>       }<br>      }<br>     }<br>    }<br>   }<br>  }<br>}</pre><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Ftenor.com%2Fembed%2F9634962&amp;display_name=Tenor&amp;url=https%3A%2F%2Ftenor.com%2Fview%2Fcomputer-monday-gif-9634962&amp;image=https%3A%2F%2Fc.tenor.com%2FU0tTlSlu5soAAAAC%2Fcomputer-monday.gif&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=tenor" width="600" height="400" frameborder="0" scrolling="no"><a href="https://medium.com/media/eeae8f0da38209953b94390ba2798edd/href">https://medium.com/media/eeae8f0da38209953b94390ba2798edd/href</a></iframe><p>We might find some issues when dealing with queries that deep:</p><ol><li>Just a single typo on the query or type definition can cause a bug that will take a lot of time to track and fix.</li><li>We might already imagine the scenario where the developer gets too frustrated keeping track of what attributes and types are in the API, especially when dealing with multi-level data objects in a single query.</li><li>As you can see, it might take a while to write all those manual types and queries, not to mention we need to edit them manually too should the query get updated (attribute change, addition, or removal).</li></ol><p>At this point, it would be great if there is something we can use to check everything for us and do all the tedious work of writing the types every time we need to integrate. So, what exactly can we do to prevent such complexity from eating up our precious work hours?</p><p><strong>Graphql Code Generator To The Rescue.</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/516/0*kfGx3xQ-cwhbhbNi" /></figure><p>GraphQL Code Generator is a CLI tool that can generate Typescript types out of a GraphQL schema and operations (<strong>query</strong>/ <strong>mutation</strong>/ <strong>subscription</strong>, and <strong>fragment</strong>). GraphQL Code Generator can be applied to the popular frontend framework, in this case, we use React, Typescript, and Apollo as GraphQL Client.</p><p>Some benefits we find when we are implementing GraphQL Code Generator:</p><ol><li>GraphQL Code Generator can generate more than just TypeScript types. It can automate some of your GraphQL development workflows, generate common practices for data fetching, and add type-safety to code you usually need to write manually.</li><li>We can generate ready-to-use React hooks for your GraphQL operations.</li><li>We can focus on the business logic that must be applied to our project.</li><li>Backend and Frontend Sync, when you make a change to the backend, the Codegen can spot any breaking changes instantly.</li></ol><p>You can start <a href="https://www.graphql-code-generator.com/docs/getting-started/installation">GraphQL Codegen Getting Started </a>and configuration for <a href="https://www.graphql-code-generator.com/docs/plugins/typescript">typescript</a>. GraphQL Codegen provides several useful plugins and presets that you can implement as needed. We’ll try to generate Schemas, and Operations using React Hooks.</p><p><strong>Generate: React &amp; Apollo hooks, and Types</strong></p><pre>// gqlconfig.js<br>module.exports = {<br> overwrite: true,<br> generates: {<br>  &#39;src/gqlcodegen/types.ts&#39;: {<br>    schema: process.env.YOUR_GRAPHQL_HOST,<br>    plugins: [&#39;typescript&#39;],<br>    config: {<br>     // ...Plugin Config<br>    },<br>  },<br>  &#39;src/gqlcodegen/operations/&#39;: {<br>    schema: process.env.YOUR_GRAPHQL_HOST,<br>    preset: &#39;near-operation-file&#39;,<br>    documents: &#39;src/gqlcodegen/operations/*.graphql&#39;,<br>    presetConfig: {<br>     extension: &#39;.ts&#39;,<br>     baseTypesPath: &#39;../types.ts&#39;,<br>     folder: &#39;../hooks&#39;,<br>    },<br>    plugins: [&#39;typescript-operations&#39;, &#39;typescript-react-apollo&#39;],<br>    config: {<br>     // ...Plugin Config<br>    },<br>  },<br> },<br>}</pre><p>At this point, GraphQL Codegen will generate our GraphQL Operations using preset ‘near-operation-file’ and plugins ‘<strong>typescript-operations</strong>’, ‘<strong>typescript-react-apollo</strong>’ from reading our query with format ‘.graphql’ inside the same destination folder.</p><p>Now we can move our queries above into the new file GraphQL Operations.</p><pre>// ~/gqlcodegen/operations/user.graphql</pre><pre>query usersWithPostsComments($options: PageQueryOptions){<br> users(options: $options){<br>  data {<br>   id<br>   username<br>   email<br>   name<br>   posts {<br>    data {<br>     title<br>      body<br>      comments {<br>       data {<br>        name<br>        email<br>        body<br>       }<br>      }<br>     }<br>    }<br>   }<br>  }<br>}</pre><p>Finally, we just need to use a magic word (or magic command, if you prefer) that can be called in the command prompt or you can register inside package.json</p><blockquote>graphql-codegen — config ./path/to/gqlconfig.js</blockquote><p>The Answers</p><blockquote>√ Parse configuration<br>√ Generate outputs</blockquote><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Ftenor.com%2Fembed%2F4533805&amp;display_name=Tenor&amp;url=https%3A%2F%2Ftenor.com%2Fview%2Fthumbs-up-computer-kid-okay-gif-4533805&amp;image=https%3A%2F%2Fc.tenor.com%2F9ItR8nSuxE0AAAAC%2Fthumbs-up-computer.gif&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=tenor" width="600" height="400" frameborder="0" scrolling="no"><a href="https://medium.com/media/8f46fe5566dc187ecffaeb958275cac7/href">https://medium.com/media/8f46fe5566dc187ecffaeb958275cac7/href</a></iframe><p>You will get a bunch of types that you can import, which is pretty handy. You also get hooks! useusersWithPostCommentQuery() and useUsersWithPostCommentstLazyQuery() which lets you pass the arguments and outcomes of your typed data.</p><pre><em>import</em> { useUsersWithPostsCommentsQuery } <em>from </em>&#39;~/gqlcodegen/hooks/user&#39;</pre><pre>const User = () =&gt; { <br>  const { data } = useUsersWithPostsCommentsQuery()<br>  <br>  <em>return</em> (<br>     &lt;div&gt;<br>      &lt;ul&gt;<br>       {data?.users?.data?.map((user) =&gt; (<br>        &lt;li&gt;<br>         {<em>/* Call user into component */</em>}<br>        &lt;/li&gt;<br>       ))}<br>     &lt;/ul&gt;<br>    &lt;/div&gt;<br>    )<br>   }<br><em>export</em> <em>default</em> User</pre><p>To get familiar, you can try generating the code for another custom query you can make from the API. This way, you will have a firm understanding of how GraphQL Codegen works, and hopefully, how it would help you work with your project!</p><h4><strong>Conclusion</strong></h4><p>In our Frontend Development team want to make our project and the code as simple and readable as possible. GraphQL Codegen can help speed up development time, and no more hassle to think about and create Types for our GraphQL Operations as GraphQL Codegen will take care of it all, so we can focus on developing business logic on every frontend project. Hopefully, you are now able to try utilizing GraphQL Codegen in your project, or at least have a ground knowledge to research further into this topic.</p><p>Please let me know in the comments if you have any questions or tips, and share if you think this article is interesting or will helpful for others.</p><p><strong>Author</strong></p><ul><li>Fajar Septiawan (Software Engineer)</li><li>Theodorus Riyanto (Software Engineer)</li></ul><p>Useful References:</p><ul><li><a href="https://the-guild.dev/blog/graphql-codegen-best-practices">Best Practices for integrating GraphQL Code Generator in your frontend applications - The Guild Blog</a></li><li><a href="https://www.graphql-code-generator.com/">GraphQL Code Generator | GraphQL Code Generator</a></li><li><a href="https://start-up.house/en/blog/articles/use-graphql-code-generator">Why You Should Use GraphQL-Code Generator with React and GraphQL</a></li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=bfb3924039ab" width="1" height="1" alt=""><hr><p><a href="https://medium.com/zebrax/utilizing-the-power-of-typescript-and-graphql-codegen-bfb3924039ab">Utilizing The Power of Typescript and GraphQL Codegen</a> was originally published in <a href="https://medium.com/zebrax">ZebraX</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Manufacturing Line Optimization using Discrete Event Simulation]]></title>
            <link>https://medium.com/zebrax/manufacturing-line-optimization-using-discrete-event-simulation-5090ecade303?source=rss----42d4c8441b9a---4</link>
            <guid isPermaLink="false">https://medium.com/p/5090ecade303</guid>
            <dc:creator><![CDATA[Barecasco]]></dc:creator>
            <pubDate>Tue, 23 Nov 2021 07:27:48 GMT</pubDate>
            <atom:updated>2021-11-23T07:27:48.301Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*MCW92xg1PDc8TVCA.png" /></figure><blockquote>The objective of this article is to describe our approach in optimizing a manufacturing line with complex process interactions using a simulation model. The approach consists of two main steps: building a simulation model, and performing optimization on the target metrics over a list of controllable variables. A simple target metrics for the optimization is the production throughput, while the controllable variables are the machine speeds or conveyor-belt sizes.</blockquote><h3>Discrete Event Simulation</h3><p>A <strong>simulation model</strong> is an imitation or replication of a real system or processes over time. The model is used to generate artificial data to help us understand the operating characteristics of the real system. When developing a simulation, it is best to start with a simple model and iterate towards a model with a balanced accuracy and complexity. After the model is ready and validated, we can then perform a what-if analysis, or run an optimization task using the model.</p><p>A <strong>Discrete Event Simulation (DES)</strong> is a modeling approach in which the state of the system changes only at discrete points in time. For this article, we are using <a href="https://simpy.readthedocs.io/en/latest/"><strong>SimPy</strong></a>,<strong> </strong>which is a process-based discrete-event simulation framework based on standard Python.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*oxzrRgICbfWwTMqP.png" /><figcaption>Figure 1. System modeling approach</figcaption></figure><h3>Manufacturing Line Abstraction</h3><p>Let’s consider a manufacturing or production line consisting of multiple machines that are connected using conveyor belts in a sequence. Each machine has a different speed (products/minute), and each conveyor belt has a different maximum capacity (number of products).</p><p>The dynamic of the manufacturing line system can be summarized as follows:</p><ul><li>The state of the machine is either running or stopped due to failure, starvation, or blockage</li><li>The machine is starved if the preceding conveyor belt is empty</li><li>The machine is blocked if the following conveyor belt is full</li><li>Machine uptime follows an exponential distribution with λ = 1/(Mean Time To Failure)</li><li>Machine downtime follows an exponential distribution with λ = 1/(Mean Time To Repair)</li><li>If historical Manufacturing Execution System (MES) data is available, the machine up and downtime can be collected and randomly sampled</li><li>At each production batch, a warm-up or a startup period will occur until the production stabilizes</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*F3KrVdEC0aY2DDRV.png" /><figcaption>Figure 2. Model abstraction in SimPy</figcaption></figure><h3>SimPy Model Setup</h3><h4>A Glimpse on SimPy</h4><p>Please note that this is a simplification of how SimPy works. Please find the documentation <a href="https://simpy.readthedocs.io/en/latest/contents.html">here</a>. There are three important terms to understand SimPy:</p><ol><li>Environment</li><li>Process</li><li>Event</li></ol><p>A simulation in SimPy runs within an Environment. In an environment, some processes need to or may not be executed. The simulation starts by prompting SimPy to execute at least one initial process:</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/def1576ba390c4a240aac5f2f661315f/href">https://medium.com/media/def1576ba390c4a240aac5f2f661315f/href</a></iframe><p>In a process, there can be event triggers and at least one event listener (the line that uses keyword <em>yield</em>). An event listener in the process will halt any code execution below it until the corresponding event is triggered. Upon the trigger, the code is executed up to another event listener or the end of the process. Event trigger can exist among the executed code so it may start code execution in another process with the corresponding event listener.</p><p>SimPy simulation runs by going from event to event. The discrete simulation time flow and simultaneity are defined by the events themselves. For example: env.timeout(3) set and listen if 3-time steps simulation can pass (not violating any other event) and it will trigger itself if the condition is met. An event trigger can be listened to by more than one event listener, meaning they will be executed simultaneously in the simulation time.</p><p>And of course, we can define our event in SimPy.</p><h4>Conveyor</h4><p>The first approach to model manufacturing lines is to model the Conveyor and the Machine.</p><p>A Conveyor object is designed to have a SimPy Container that can be loaded or unloaded. The container can represent the buffering function of the conveyor belt. In the real world, it represents how items can crowd up on the belt.</p><p>SimPy has a built-in class called Container that can be used to put and take several items from it. It has the capacity where capacity-related constraint definition can be applied. Below is the instantiation of the Container class within the constructor function of the Conveyor class.</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/c3bd7f30f48d76d4a594a4942aa91936/href">https://medium.com/media/c3bd7f30f48d76d4a594a4942aa91936/href</a></iframe><p>Loading the container will take some time, while unloading happens instantly. The time taken to load the container represents the time it takes for an item to travel from the start to the end of the conveyor belt. Instant unloading represents how the next machine takes in items at the end of the conveyor belt. Controlling how the time flow for loading and unloading in the simulation is done via timeout event. The speed of the conveyor belt is represented by the number of time steps used in timeout.</p><p>There are two conditions regarding the container capacity. When the preceding machine attempts to put items to the container while the result will exceed its capacity (container.level + num_item &gt;= container.capacity), a <em>blockage</em> event is triggered so the preceding machine will stop its process. When the next machine attempts to take in items from the container while it has no sufficient number of items (Container.level - num_item &lt; 0), a <em>starved</em> event is triggered so the next machine will stop its process.</p><p>Below is an example of a load method of Conveyor class that implements conveyor belt traveling time and over-capacity condition:</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/6f00f1a0ec4699e06d108270d7d044bd/href">https://medium.com/media/6f00f1a0ec4699e06d108270d7d044bd/href</a></iframe><p>Once the container is ready to be loaded or unloaded, the event <em>load_ready</em> or <em>unload_ready</em> is triggered so the corresponding machine can resume its process.</p><h4>Machine</h4><p>A Machine object is designed to be connected with input and output conveyor instances. The connection will allow the machine to do a request to the conveyors’ container to put and get items from it.</p><p>Once the machine takes in an item it will be processed for some time steps. The amount of time steps for thetimeout event is determined by desired products/minute speed parameter and time step equivalent in the real world. After the processing is done the item will be put to the output conveyor.</p><p>The connected conveyor can trigger blockage or starved event if its capacity rule is violated on loading or unloading attempt. A stopped machine will wait until the conveyor is available to be loaded/unloaded, represented by <em>load_ready</em> or <em>unload_ready</em> event from the corresponding conveyor.</p><p>Below is the example of how a Machine class with blockage event logic can be defined:</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/dde864d919920008a20a61fc9a1aa277/href">https://medium.com/media/dde864d919920008a20a61fc9a1aa277/href</a></iframe><h4>Machine Failure</h4><p>A machine failure (or breakdown) is simulated as a periodic event where the time between two failures is generated from a random number following an exponential distribution. A mean time between failure is required. In SimPy we can create a process that will interrupt the machine when the breakdown time has come.</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/c3b81e3c8d77e30527682d19879a0079/href">https://medium.com/media/c3b81e3c8d77e30527682d19879a0079/href</a></iframe><p>The same logic applies to simulate the repairing process of the machine. The mean time to repair is required to generate the exponentially distributed random number for repair time.</p><h4>Simulation Arrangement</h4><p>Once the Machine and Conveyor object is defined, we can arrange the simulation objects and parameters. Below is an example of a manufacturing line consisting of 2 machines connected to 3 conveyors.</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/e0b1f59b9e7ad89d2c48f1dc18aac4f2/href">https://medium.com/media/e0b1f59b9e7ad89d2c48f1dc18aac4f2/href</a></iframe><p>The above arrangement will run up to 200 simulation time steps. For real use cases, more machines and much more time steps are required. The result of the simulation can be obtained by logging the system state for every event and computing the throughput for the whole given simulation time.</p><p>For the optimization purpose, for every machine and every conveyor, we can set its speed and items capacity within a predetermined boundary.</p><h3>Optimization Process</h3><p>One thing to note is that the solution found is not guaranteed to be global optimum due to the non-linearity of the problem. However, we can still validate the throughputs between the current line configuration, and the optimized configuration to assess the performance improvement.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*lzBrNLAPLvgTJmpb.png" /><figcaption>Figure 3. Optimization process using the black-box simulation model</figcaption></figure><p>Results from the optimization task can be visualized in terms of the v-shape graph to understand the current bottleneck and be used as an input to improve the line performance by adjusting the machine speed or the conveyor-belt capacity.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*RrVPa-wGa3l4dcyx.png" /><figcaption>Figure 4. Manufacturing line v-shape graph</figcaption></figure><h3>Summary</h3><p>In this article, we’ve covered the general idea of manufacturing line optimization or what is commonly known as line balancing analysis. The key benefit of this solution is to provide line managers a tool to simulate production lines, perform what-if analysis, and optimize throughput.</p><h3>Author</h3><ul><li>Agra Barecasco (Sr. Data Scientist)</li><li>Edward Christian (Data Architect)</li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=5090ecade303" width="1" height="1" alt=""><hr><p><a href="https://medium.com/zebrax/manufacturing-line-optimization-using-discrete-event-simulation-5090ecade303">Manufacturing Line Optimization using Discrete Event Simulation</a> was originally published in <a href="https://medium.com/zebrax">ZebraX</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Bas]]></title>
            <link>https://medium.com/zebrax/bas-3c269a3a5fcd?source=rss----42d4c8441b9a---4</link>
            <guid isPermaLink="false">https://medium.com/p/3c269a3a5fcd</guid>
            <category><![CDATA[tips]]></category>
            <category><![CDATA[ui-design]]></category>
            <category><![CDATA[ui-ux-design]]></category>
            <category><![CDATA[enterprise-ux]]></category>
            <category><![CDATA[ux-design]]></category>
            <dc:creator><![CDATA[zxdesign]]></dc:creator>
            <pubDate>Fri, 22 Oct 2021 08:51:22 GMT</pubDate>
            <atom:updated>2021-11-26T09:48:41.127Z</atom:updated>
            <content:encoded><![CDATA[<h3>Basic Tips to Enterprise UX</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*UC8FYRuL9Uh_hVuXle-gqw.png" /></figure><p>As a designer we studied a lot about “keep it simple”, “less is more” and “don’t make me think” concept, those are a great concept when creating an app, but how do we use those principles when creating an Enterprise Apps, or Apps for Industries, those principles can sometimes be detrimental to good user experience. Simplicity can clash with the fundamental complexity of enterprise apps and processes.</p><p>In short Enterprise, Apps are Internal Apps used by Enterprises / Industrial to fulfill their needs. A complex app used by their employee, mostly containing a chart, table, form, etc.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*cLWgnAutcg5QoU8O1swpyw.png" /></figure><p>Keep in mind that we often find it difficult to implement those most popular principles above when it clashes with what our user needs.</p><p>Enterprise apps are often too complex with many features and a lot of data or we call it “feature creep”, we can’t fully avoid it because it can affect their business process involving a lot of our customer money.</p><p>We might be tempted to implement our ideal solution on the platform, then we wonder is it the best way to present it.</p><p>With that in mind, how do we make sure designing these complex features apps help them achieve their goal? While enterprise applications are generally much more complex than consumer applications, a UX designer should minimize the impact of that complexity on users, here’s some way to achieve that.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/239/1*STV3jMZJTZY-4axmRX8g-g.png" /></figure><p>Thank you</p><h4>Creating a Role-Based UI</h4><p>To make the system as simple as possible, though never as simple as we mean to be, creating a role-based UI can be helpful to revealing feature functionality only when a user needs it. So particular users see only the functionality they need. Often, achieving simplification means battling features to eliminate functionality that users don’t need or better, to avoid adding it in the first place.</p><h4>Understanding The Users</h4><p>Unlike the usual customer-facing apps, when we’re working on an enterprise application, we’re designing for people who are likely to use it day in, day out. They learn the parts of it that are most relevant to their particular work.</p><blockquote>“ One of the biggest differences between enterprise and consumer applications is that most users of enterprise applications don’t have a choice about using them, their employee telling them to do so “</blockquote><p>Extra empathizing with our users and understanding their pain points can help us make a better decision for the app’s design. Qualitative research can be an effective way to collect data from our users. We can create an interview session or Forum Group Discussion. Having an actual insight from our specific user can help us make accurate decision making.</p><h4>Complex but not Complicated</h4><p>Simplification doesn’t always work in an enterprise environment. As an outsider or not expertise, we can’t oversimplify the data or information that our user needs. It can create misleading information and prevents professionals from doing their job successfully.</p><p>Enterprise applications tend to offer more features to provide business intelligence, workflows, collaboration, and productivity-management capabilities. So enterprise applications still tend to be more complex than consumer applications. This often leads UX designers to use more hide-and-peek design patterns. The intent of these patterns such as tabs, accordions, and carousels is to make large data sets or amounts of content more digestible.</p><h4>Understand the Business Model and Goals</h4><p>The first thing to remember is that we’re designing for busy professionals who need to complete their tasks. The business goal is not the same with ease of use. Tried to understand our user’s goals, no matter how hard or unfamiliar you are in their industries or expertise.</p><p>Always check all documents related to your project, be it business documents or data documents. It can help you when designing your platform whether those documents need to be shown in actual number or percentage number. For example, some companies show their goal achievement with percentage number and the other use a comparison between goals achieved and remaining goals.</p><h4>Use the Right Chart</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/482/1*ngtrxwJaMW8C-47-aRbh1w.png" /></figure><p>The main function of a chart is to display data. When making your chart, think about the specific information that you want your data to support or the outcome that you want to achieve.</p><p>Keep your chart simple, bombarding an audience with data will likely leave them confused and uncertain. A basic bar chart or pie chart can be the best practice to show the visualization of data.</p><h4>Conclusion</h4><ul><li>Creating a Role-Based UI can be helpful to reduce distraction and mitigate mistakes in users’ daily activity.</li><li>Extra empathizing to the user is needed, they’re going to use the system 7–5 or even more every day they working.</li><li>Understanding the Business Model and Goals can help make a correct decision.</li><li>Don’t oversimplify, it prevents professionals / user from doing their job properly.</li><li>Make users navigate to what they need quickly.</li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=3c269a3a5fcd" width="1" height="1" alt=""><hr><p><a href="https://medium.com/zebrax/bas-3c269a3a5fcd">Bas</a> was originally published in <a href="https://medium.com/zebrax">ZebraX</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Story of a Lazy Developer]]></title>
            <link>https://medium.com/zebrax/be-an-efficient-developer-76ee9ab2656b?source=rss----42d4c8441b9a---4</link>
            <guid isPermaLink="false">https://medium.com/p/76ee9ab2656b</guid>
            <category><![CDATA[application-developer]]></category>
            <category><![CDATA[developer-stories]]></category>
            <category><![CDATA[application-development]]></category>
            <dc:creator><![CDATA[ZebraX Engineering]]></dc:creator>
            <pubDate>Fri, 15 Oct 2021 08:51:29 GMT</pubDate>
            <atom:updated>2021-10-15T10:03:18.502Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*UG10-AtS-HghUBMHqAWHDw.png" /></figure><p>Have you ever met a software developer with messy hair? Or have you ever met some that never take a bath? Or are you one of them? Congratulations! You’ve met the requirement to be a good developer. Remember when Bill Gates said that he only chooses a lazy person to do a hard job. Because a lazy person will find an easy way to do it. Lazy people tend to do something efficiently because they don’t have time to do unimportant things.</p><h3>The perks of being a lazy developer</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/640/1*vjV3jmqo3O67UBhsTJRgqw.png" /></figure><p>Being lazy doesn’t mean doing things poorly. It’s the opposite. A lazy developer would not create a hard-to-maintain code, undocumented code, or spaghetti code. Because they know that sometime in the future they will go back again to that code. If they created the code in the best way possible. They will not waste their ‘precious’ time to re-study their code in the future.</p><p>Lazy developers know how precious every code that they make is. Hence, they created the code as simple and elegant as possible. They will not reinvent the wheel but also avoids using unnecessary frameworks and libraries.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/640/1*rQuJG7Yujd99IY6VmChlWA.png" /></figure><p>Lazy developers only focus on important things. Because they value their time, they work only on necessary things. Sometimes doing nothing is better to focus your mind more. Because the hardest part of being a software developer is thinking.</p><h3>Automate Repetitive Task</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/640/1*iI2Jw9B92_cdNgSDE09FhQ.png" /></figure><p>Lazy developers hate doing repetitive tasks. They easily get bored and want to end things quickly. One way to achieve this is using automation. There is some saying “you could have done the task in 6 minutes but instead, you automated it for 6 hours”. I think that’s only partially true. Because that 6 hours you make for automation will save you repetitive 6 minutes. Which if we sum that repetition, it could be more than 6 hours. And if you’re lazy you know which one’s better.</p><h3>Examples of Lazy Programmers</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/640/1*AIG-kCAT66TSYpPF0E2ldQ.png" /></figure><p>Take an example in ZebraX especially on the front-end side. We implement Prettier together with Eslint to automate the formatting process for our code. We don’t have to think about the spacing, the indentation, the forgotten variables that we don’t use, the forgot-to-remove ‘console.log’ and so much more when we finished coding. Prettier and Eslint will do the job for us. Imagine if we do this manually. There will be so much time wasted that could be used for more important things.</p><p>We also have our own CI/CD pipeline with Jenkins, so we don’t need to spend time deploying our code manually to our development, UAT, or production server. Every time we merge Pull Request on our Git repository, the code will be built and deployed automatically on our environment. In that way, we only need to focus on our code, no need to go into our server and deploy it manually every time we have changes in our code.</p><p>Not only that, but we also use SonarQube integrated with Jenkins in our GitHub projects. Every time we make a Pull Request, SonarQube will automatically review our code. SonarQube will detect duplicated code, coding standards, unit tests, code coverage, code complexity, comments, bugs, and security vulnerabilities. This also comes in handy because sometimes reviewing code can be a little bit frustrating and it’s very possible to miss it if it’s done manually.</p><p>References:<br>1. <a href="https://www.parkersoftware.com/blog/hire-lazy-developer/">https://www.parkersoftware.com/blog/hire-lazy-developer/</a><br>2. <a href="https://betterprogramming.pub/the-lazy-mindset-of-an-effective-programmer-68f4b70e3aac">https://betterprogramming.pub/the-lazy-mindset-of-an-effective-programmer-68f4b70e3aac</a><br>3. <a href="https://dev.to/drm317/the-lazy-developer-5goe">https://dev.to/drm317/the-lazy-developer-5goe</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=76ee9ab2656b" width="1" height="1" alt=""><hr><p><a href="https://medium.com/zebrax/be-an-efficient-developer-76ee9ab2656b">Story of a Lazy Developer</a> was originally published in <a href="https://medium.com/zebrax">ZebraX</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 Efficiently Train Multiple ML Models on a Spark Cluster]]></title>
            <link>https://medium.com/zebrax/how-to-efficiently-train-multiple-ml-models-on-a-spark-cluster-7d84512d36f0?source=rss----42d4c8441b9a---4</link>
            <guid isPermaLink="false">https://medium.com/p/7d84512d36f0</guid>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[spark-cluster]]></category>
            <category><![CDATA[multiple-ml]]></category>
            <category><![CDATA[zebrax]]></category>
            <dc:creator><![CDATA[ZebraX Team]]></dc:creator>
            <pubDate>Wed, 15 Sep 2021 09:51:34 GMT</pubDate>
            <atom:updated>2021-09-15T09:51:34.027Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*hmHPub6WpgfGMQzHDTsFkw.png" /></figure><h3>Introduction to Spark</h3><p>Spark is a fast, distributed analytics computing engine for large-scale data processing and machine learning modeling. Spark allows data sharing between processing steps through<strong> in-memory processing</strong> of data pipelines, hence it can run workloads a lot faster if compared to Hadoop. Spark also comes with powerful libraries that can be used together in the same application: SparkSQL, SparkML, SparkStreaming, and SparkGraphX.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/930/0*twO4hbwKPZ077_9z" /></figure><p>SparkML provides multiple types of machine learning algorithms, including classification, regression, clustering, and collaborative filtering, as well as supporting functionality such as model evaluation and data import. It is designed to perform machine learning algorithms on<strong> distributed data</strong>.</p><h3>Training a Multiple Partitioned Models</h3><p>Imagine if we are asked to build a different model for each partition in our dataset, e.g. product types, regions, etc. Let’s say we have 100 different partitions, training 100 models in a serial mode will be time-consuming and not efficient. Parallel modeling can provide tremendous benefit in building a variety of models by speeding up the process through parallelization so the model building process becomes more efficient.</p><p>Building a machine learning model in Python, using its popular and powerful libraries such as Scikit-learn, Pandas, and Numpy, is easy and fun. But Python has its limitations when it comes to large-scale data processing. Python libraries are not set up to work in a distributed environment and are bound on a single compute machine with one contiguous block of memory, which make it infeasible to be used for training on large-scale distributed dataset.</p><p>With the help of <strong>Spark Pandas UDFs</strong>, we can partition our dataset by the specified category (column) into smaller chunks, then Spark will distribute each chunk across the cluster nodes and execute the Python code on each node. This feature also allows us to bring many Python functionalities from its massive libraries such as SciPy, NumPy, Scikit-learn, and StatsModels, just to name a few. Now, with Pandas UDFs there is a new scale of compute that is available for Python developers.</p><p>In order to bring powerful functionalities of Python’s libraries into Spark cluster framework, we need a new feature that enables us to run our Python code in a distributed environment. Fortunately, Databricks introduced the Pandas UDFs (a.k.a. Vectorized UDFs) feature in the Apache Spark 2.3 release that substantially improves the performance and usability of user-defined functions (UDFs) in Python.</p><p>Instead of pulling the full dataset into memory on the driver node, we can use <strong>Pandas UDFs</strong> to distribute the dataset across a Spark cluster, and use <strong>Apache Arrow</strong> to translate between the Spark and Pandas data frame representations.</p><h3>Pandas UDF Concept</h3><p>Pandas UDFs are features that are executed by Spark which enable Python code to run in a distributed environment. User-Defined Functions or UDFs use Apache Arrow to transfer data and Pandas to work with the data <a href="https://docs.databricks.com/spark/latest/spark-sql/udf-python-pandas.html">[1]</a>. There are two types of Pandas UDFs, scalar and grouped map UDFs. <strong>Scalar</strong> is used for vectorizing scalar operations while <strong>grouped map</strong> is designed for splitting combine pattern analysis based on conditions specified in the groupby operator <a href="https://databricks.com/blog/2017/10/30/introducing-vectorized-udfs-for-pyspark.html">[2]</a>.</p><p>For using Pandas UDFs, both scalar and grouped map are adding the @pandas.udf decorator above the function. But, they have some differences like input and output for scalar is Pandas series while grouped map is Pandas dataframe, scalar has no grouping semantics while grouped map defined by groupby clause, and the return types for scalar is a DataType and grouped map is a StructType.</p><ul><li>Grouped Map Feature</li></ul><p>ZX AnalytiX application uses the grouped map feature, which was initially introduced in Spark 2.3. In summary, this feature allows us to partition a Spark dataframe into smaller chunks that are converted to Pandas dataframes before being passed to our Python code. UDFs are executed on worker nodes, enabling existing Python code to now be executed in a distributed manner.</p><p>​​However, a few conditions need to be met for a problem to match this use case. The first prerequisite is that our data can be well partitioned by a categorical column (key), and the second prerequisite is that our data needs to be represented as a single Spark dataframe. The grouped map feature is able to distribute a single dataframe across worker nodes using a partitioning column, and only supports the groupby operation on a single object.</p><ul><li>Inherent Risks of using pandas UDFs</li></ul><p>The first issue is that debugging Pandas UDFs was much more difficult than working with Pandas dataframes on the driver node, because we can no longer use print statements to trace execution flow. Current best practice for testing Pandas UDFs is to first use the toPandas method on a small dataframe and write a groupby apply function that runs on the driver node before trying to distribute the operation for the full dataset.</p><p>The second issue is that the schema for the returned Pandas dataframe needs to be specified before the function is defined, since the schema is specified as part of the grouped map annotation. With this, you cannot store your trained model objects directly to the dataframe. You need to convert your model into a string representation first and then store it in the dataframe, this link <a href="https://patrickroos.org/2019/02/25/training-many-scikit-learn-models-using-pyspark-and-pandas-udfs/">[3]</a> gives you a clear example of how to do that.</p><p>Another trick you can do is to store your trained models into a local/cloud storage. By doing this you can easily manage your models, evaluate, and use it whenever we want to make predictions on new data.</p><h3>Parallel Modeling using Spark Pandas UDF</h3><p>When working on a parallel model training, which aims to parallelize the training of machine learning model processes, the suitable Pandas UDFs type is grouped map Pandas UDFs. With this, we can build a machine learning model on each partition using the <strong>split-apply-combine pattern</strong> that grouped map UDFs offer.</p><p>The steps are simple, first is defining a Python function for the training process, then put the pandas udf decorator to it, so it will enable the parallelization. Then define its output schema, and the function as a<strong> PandasUDFType.GROUPED_MAP</strong>. That’s it, now we are ready to apply this Pandas UDF to our dataset.</p><p>Applying the Pandas UDF is also simple. The most important part that we need to do first is to ensure that our dataset can be well divided by a categorical column that we are interested in. If it is not, then we should do data conditioning before applying the function.</p><p>The next step is to use the groupby method to your dataframe, specify the partition column, and apply the Pandas UDF. Spark will execute the Pandas UDF on these smaller chunks of pandas dataframe and combine all the results back to one Spark dataframe.</p><h3>Quick Demo Using ZX AnalytiX</h3><p>Let’s take a look at the ZX AnalytiX platform below. In this platform we already implement the Pandas UDF grouped map feature to scale up the application of parallel modeling. Using this platform, we can build a model pipeline and then train the model without doing any coding, which is great to speed up our work in model building.</p><p>In this simple demo, we use toy data of activities of assets in a mining area. The asset types are car and wheel loader. These two assets have very different behaviors/activities, so it is a right choice to build a different model for these two assets. The goal of our model is to predict the asset status, whether it is <strong>idle, active, </strong>or<strong> off</strong>.</p><p>Now, we already understand our problem and goal, then we can continue to build our model pipeline. It starts with choosing the data from the database, choosing a partition column (the asset_type), and choosing feature engineering, feature selection, and feature handling methods if necessary.</p><p>The final configuration of our pipeline is to choose the algorithm(s) that we want to use for our models. We can choose several algorithms for this training phase. The application will pick over the best model for each partition based on the user specified evaluation metric.</p><p>After we execute the models, the following result shows that there are 3 best models for each partition (two models represent model for each partition, and one model represents model without partition) that were obtained from the parallel model training process.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*GAqmqB-qO3c8zcZy" /><figcaption><em>ZX AnalytiX Partition Model — Experiment Report</em></figcaption></figure><p>Below is a short clip of the ZebraX DMAA Platform for this simple use case</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*uyf_WXNXXAcsvuQuMW260Q.gif" /></figure><h3>Summary</h3><p>Pandas UDF is a great feature that enables us to run our Python codes in a distributed environment. We showed that we can benefit from this functionality to build multiple models in parallel manner, but the application of Pandas UDF is not limited to model training. It can be used for feature generation, statistical testing, and other variety of applications for data science.</p><blockquote>At ZebraX we have used Pandas UDFs to improve our application capabilities, so that we can produce greater data-driven insights to our customers. By leveraging Spark capabilities for distributing the computing process and powerful functionalities of Python’s libraries, we aim to add more capacities to our Advanced Analytics platform to tackle diverse business problems using data-driven approaches.</blockquote><h3><strong>Author</strong></h3><ul><li>Gilang Samudra (Senior Data Scientist)</li><li>Puteri Aulia Indrasti (Data Engineer)</li></ul><h3>References</h3><ol><li><a href="https://docs.databricks.com/spark/latest/spark-sql/udf-python-pandas.html">https://docs.databricks.com/spark/latest/spark-sql/udf-python-pandas.html</a></li><li><a href="https://databricks.com/blog/2017/10/30/introducing-vectorized-udfs-for-pyspark.html">https://databricks.com/blog/2017/10/30/introducing-vectorized-udfs-for-pyspark.html</a></li><li><a href="https://patrickroos.org/2019/02/25/training-many-scikit-learn-models-using-pyspark-and-pandas-udfs/">https://patrickroos.org/2019/02/25/training-many-scikit-learn-models-using-pyspark-and-pandas-udfs/</a></li></ol><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=7d84512d36f0" width="1" height="1" alt=""><hr><p><a href="https://medium.com/zebrax/how-to-efficiently-train-multiple-ml-models-on-a-spark-cluster-7d84512d36f0">How to Efficiently Train Multiple ML Models on a Spark Cluster</a> was originally published in <a href="https://medium.com/zebrax">ZebraX</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Intern’s Story: Journey to be a Software Engineer]]></title>
            <link>https://medium.com/zebrax/interns-story-journey-to-be-a-software-engineer-597e4f4c39dc?source=rss----42d4c8441b9a---4</link>
            <guid isPermaLink="false">https://medium.com/p/597e4f4c39dc</guid>
            <category><![CDATA[internship-experience]]></category>
            <category><![CDATA[software-engineer-life]]></category>
            <category><![CDATA[zebrax]]></category>
            <dc:creator><![CDATA[Theodorus Riyanto]]></dc:creator>
            <pubDate>Wed, 25 Aug 2021 05:27:01 GMT</pubDate>
            <atom:updated>2021-08-25T05:27:01.243Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*zCLu_uWVsKgl0vthrmzzWg.png" /></figure><p>Hello there! My name is Theodorus Riyanto, and I would like to share my interesting intern story at <a href="http://zebrax.id/">ZebraX</a>. A little bit of background, before I started my internship, I was still a student in Institut Teknologi Bandung, majoring in Petroleum Engineering. I can already imagine you reading this while wondering, why a petroleum engineering student, whose knowledge all about working in an oilfield, and not to mention is in his last semester, would bother looking for a software engineering internship? Well, I was just as surprised, but hang on as I’m about to spill a little bit of how I got to that point.<br> <br> Not to deviate too far from my intern story, my interest in software engineering sparked when I was in college. Definitely not when I enter ITB (because if so, I would have picked a different major), nor it was when I entered my last semester as a panic diversion from my current major. It was in my first year in ITB, that I had the chance to learn about programming, in a mandatory “introduction to programming” class. At that time, it was not even software engineering, but more about algorithm and data structure that got me thinking “Wow, maybe I should explore deeper about this programming thing”. Later I was recruited as an assistant for that particular class, which was the most enjoyable experience I ever had with programming at the time.<br> <br> Fast forward to my last semester, I had the chance to revisit programming since I have more free time and was not really satisfied with what I have explored so far. I have previously been working in petroleum industry and business as an intern, but it wasn’t clear for me which career path to pursue. And there was my programming experience, sitting idle, waiting for me to pick it up and see if it suits me better. Since I still have more time to spend, especially during the pandemic where remote working is a thing, I decided to give this programming thing a shot and start aiming to get a Software Engineer Intern as the starting point.<br> <br> Around March, after spending months of learning, practicing, and building projects, I managed to land an internship at ZebraX, which is quite an opportunity for me considering the few skills I have and no prior internship experience in the field, but I believe they saw something in me and decided to have me as a software engineering intern.</p><p>Since it was my first internship, I expected to learn a lot and require a lot of assistance in understanding how everything works together and how I could contribute.</p><blockquote>Luckily, I was put in one of their front-end engineering team, and my mentor was very capable of guiding me through all the questions that I had, so I could pick up faster and be productive as soon as possible. There was definitely a lot of efforts that I had to put to be able to learn so many things, but having the right mentor definitely makes it easier.</blockquote><p>I remember the first day of my internship I was trying my best to learn as fast as possible while contributing to the project, and at the end of the day, I managed to push my first Pull Request (PR), even though it was not much for an 8-hour effort. I was the kind of person to read and search first before I ask, because I believe the process it takes to figure out something is very important to my growth as a software engineer. The work culture provides me an opportunity to do so, and that was exactly the approach I took during my internship. I could proudly say, that single PR at my first day had grown in numbers, resulting in a lot more contributions as I get better and better.</p><blockquote>During my internship at ZebraX, I was involved in several projects that shaped me into a better developer, since I gain exposure to different styles and implementations. After my three-months internship, I definitely see a noticeable growth since the time I started my internship. I am very grateful for my kind and exceptional mentors, my supportive leads and managers, and also the friends I have made along the way that gives me assistance and keeps me on my feet. Thanks to ZebraX, now I have more confidence to walk down the road of Software Engineering!</blockquote><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=597e4f4c39dc" width="1" height="1" alt=""><hr><p><a href="https://medium.com/zebrax/interns-story-journey-to-be-a-software-engineer-597e4f4c39dc">Intern’s Story: Journey to be a Software Engineer</a> was originally published in <a href="https://medium.com/zebrax">ZebraX</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Optimizing Mobile Experience in Asset Productivity]]></title>
            <link>https://medium.com/zebrax/optimizing-mobile-experience-in-asset-productivity-c8821e0736a3?source=rss----42d4c8441b9a---4</link>
            <guid isPermaLink="false">https://medium.com/p/c8821e0736a3</guid>
            <category><![CDATA[zebrax]]></category>
            <category><![CDATA[mobile-experience]]></category>
            <category><![CDATA[digital-transformation]]></category>
            <category><![CDATA[ui-ux-design]]></category>
            <dc:creator><![CDATA[zxdesign]]></dc:creator>
            <pubDate>Fri, 25 Jun 2021 04:31:48 GMT</pubDate>
            <atom:updated>2021-06-25T04:31:48.839Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*hjFsdNAogY1ukmdiFmrurw.png" /></figure><p><em>UX design isn’t a linear process. To make great products, you need to make lots of changes and test them. The key is to define assumptions, test them, refine, and repeat.</em></p><p>As a business, we constantly challenge ourselves to provide the most innovative solutions to our clients. There are changes and updates to the project to provide the most suitable solutions, and it is first mirrored through the User Interface (UI). We test, discuss and iterate new changes, and it goes constantly on and on to achieve a great User Experience (UX). A good UI/UX might allow users to be able to navigate the program, but a better one will allow users to make use of the full potential of the overall project.</p><p>In<a href="http://zebrax.id/"> ZebraX</a>, we develop an Asset Productivity Dashboard. The Asset-Intensive Industries — such as chemicals, oil and gas, mining, metals, pulp and paper, and power production — have been turning into new technologies to increase the reliability and availability of their equipment. The technology helps each company to keep maintenance costs under control, where users can track the asset activity, location and history. It is a real-time single source of truth of asset productivity, fast and easy to convey information to the whole organization.</p><p>Having real-time insight in terms of asset operations helps to monitor the business, define areas for improvement, and conduct accurate decision making. And as we completed the Asset Productivity Dashboard to help the companies, we tested and iterated the design once again, and we decided to work through it on a different screen, the mobile.</p><blockquote>Creating the mobile screen at the end of our design process is challenging. Occasionally, the design is made mobile-first rather than desktop-first, because it is a small screen that only can contain so little information. Mobile-First design helps the design process to focus on the most important content first in a hold of the hand. It is a design strategy to sketch, prototype and work on your way into the smallest screen then way up to the largest. This strategy helps to create a consistent user experience across all devices.</blockquote><p>As “Desktop First” is not a common practice, we’re up for a challenge to create the mobile at the end of our design process. It is designed considering the asset tracking process that not only happens in the room, but also on the site.</p><h3>Creating Mobile Asset Productivity</h3><h4><strong>Content First</strong></h4><p>As experts say, desktop to mobile is a “Graceful Degradation” but we make it happen</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/841/0*YQ8qBkrsnlI3yWtS" /></figure><p>We built a desktop version with well-rounded features but then needed to cut some of the features. We need to prioritize the features, knowing the information architecture of the project is the key. Which content is the most important to this project? To work on the smallest screen means to focus on the most important metrics and focus on it.</p><p>In this project, the main metrics are the activity of the asset, where we can see the asset is either on and working, on but idle, or off. This metric affects fuel consumption, staffing decisions, maintenance scheduling, and rental costs. Understanding exactly how an asset is being used is the first step to controlling costs. Making the design fit the hold of the hand is challenging but as we know the main metrics, we managed to fit them in.</p><h4>Simplicity and Straightforward</h4><p>As we understand that activity is the priority, we created the page around it. The main focus of the mobile page is a donut chart, where we could see the largest or the smallest percentage of the activity. Users can easily translate which activity is the largest and define areas for improvement of the current asset.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*Ya8d8qPVM8h42GDQ" /></figure><h4>User Friendliness</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/618/0*zArrKDTk2rYAN1_Y" /></figure><p>The mobile version does not need complicated features or components that are hard to navigate. As we put the main content which is on the first page, we created a simple tab on top of it for the users to navigate between the dashboard and the asset location tracking. Users can see the asset location on the map and also go to the asset details as we create a details button on the map.</p><blockquote>UX design isn’t a linear process. To make great products, you need to make lots of changes and test them. The key is to define assumptions, test them, refine them, and repeat them. We test and refine our Asset Productivity Dashboard to provide the most effective solutions to the users. We hope to deliver excellent information through Mobile Asset Productivity Dashboard and successfully communicate the data to help monitor the business.</blockquote><blockquote>Changes are inevitable, and we’re up for the challenge.</blockquote><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c8821e0736a3" width="1" height="1" alt=""><hr><p><a href="https://medium.com/zebrax/optimizing-mobile-experience-in-asset-productivity-c8821e0736a3">Optimizing Mobile Experience in Asset Productivity</a> was originally published in <a href="https://medium.com/zebrax">ZebraX</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>