<?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 Teigen Olson on Medium]]></title>
        <description><![CDATA[Stories by Teigen Olson on Medium]]></description>
        <link>https://medium.com/@teigenolson?source=rss-c7e664d75f83------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*dmbNkD5D-u45r44go_cf0g.png</url>
            <title>Stories by Teigen Olson on Medium</title>
            <link>https://medium.com/@teigenolson?source=rss-c7e664d75f83------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Wed, 18 Nov 2026 08:18:58 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@teigenolson/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[A Field of 0’s and 1’s: A Look at Data Science in Agriculture]]></title>
            <link>https://medium.com/@teigenolson/a-field-of-0s-and-1-s-a-look-at-data-science-in-agriculture-a88cc12ae2d4?source=rss-c7e664d75f83------2</link>
            <guid isPermaLink="false">https://medium.com/p/a88cc12ae2d4</guid>
            <category><![CDATA[yield-farming]]></category>
            <category><![CDATA[agriculture]]></category>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[agriculture-technology]]></category>
            <category><![CDATA[data]]></category>
            <dc:creator><![CDATA[Teigen Olson]]></dc:creator>
            <pubDate>Tue, 09 Nov 2021 06:45:20 GMT</pubDate>
            <atom:updated>2021-11-17T14:12:26.337Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1000/1*ysLTEDvUgYrJCIxz-U_05g.jpeg" /></figure><p>In the age of the technological revolution, we are seeing the implementation of data analytics and computer science in general in many industries around the world. One industry that I am particularly interested in for a few reasons is the agricultural industry. For one, it feeds the world, however, it harms the planet in doing so because of poor practice so it’s kind of an interesting double edged sword, providing us with an ‘in between a rock and a hard place’ sort of situation, where our only way out could be… the use of data science. The other reasons why it interests me is because of my education in Agricultural and Foods Business Management as well as my dad’s childhood.</p><p>My dad grew up on a farm in rural western Minnesota. It was primarily operated by his father and himself until my dad moved away from home after college. Because of the vast amount of time he spent planting, harvesting, taking care of animals, and everything in between, I have the pleasure of hearing of numerous recounting from his childhood. Hard work and early mornings/late nights are a common theme among these stories, as well as stories about the harder years. Hearing about these struggles in the industry and hearing about problems in agriculture while learning about data science has sparked my interest regarding the possibilities of utilizing data science in agriculture to increase production efficiency while cutting back on waste.</p><p>One scholarly <a href="https://www.mdpi.com/1424-8220/18/8/2674/htm">article</a> that I came across investigated the current and possible uses of data science in the agricultural industry while focusing on “(a) crop management, including applications on yield prediction, disease detection, weed detection crop quality, and species recognition; (b) livestock management, including applications on animal welfare and livestock production; water management; and (d) soil management”. However, because most of my dads stories focus on being out in the fields, I am going to focus on explaining how data science has been used in crop management.</p><p>Crop Management</p><p>Data science can and has been used in multiple applications within the crop management sector including in yield prediction, disease/weed detection, crop quality, and species recognition. There have been different models and algorithms used for each and to display which have been used in each sector, I have provided a list for each below.</p><blockquote>Yield Prediction</blockquote><p>Artificial/Ensemble Neural Network, Expectation Maximization, Support Vector Machines,</p><blockquote>Disease/Weed Detection</blockquote><p>Support Vector Machines, Artificial Neural Networks, Counter-propagation Neural Network</p><blockquote>Crop Quality</blockquote><p>Support Vector Machines, Ensemble Learning, Random Forest</p><blockquote>Species Recognition</blockquote><p>Convolutional Neural Networks, Deep Learning</p><h3>So, what are all of these algorithms and models?</h3><p>Neural Network: a network of ‘nodes’ which make it possible for the computer to learn.</p><p>Expectation Maximization: an iterative method algorithm used to find the maximum possibility or estimates for parameters in a model.</p><p>Support Vector Machines: a supervised machine learning technique which attempts to form a hyperplane in the nth dimension that classifies all of the data points.</p><p>Random Forest: an ensemble technique which uses multiple decision tree and then uses average outputs from various tree to form a prediction.</p><p>Ensemble Learning: This is a model which uses multiple algorithms to obtain a better result than would have been possible if just using a single algorithm.</p><p>Deep Learning: this method attempts to mimic the human brain by clustering information and using this to make predictions. Essentially a neural network with three or more layers. Commonly used in artificial intelligence applications.</p><h3>How did these algorithms help agriculture?</h3><p>In yield prediction, one study looked at predicting the yield of wheat fields using an Artificial Neural Network and did so with 81.65% accuracy. Another study did this for all crops in general using Artificial and Ensemble Neural Networks with a 1.3% error rate. Successful models in yield prediction is important because it can help farmers get a better understanding of which factors will be the most important during the planting and growing periods of the season.</p><p>In disease and weed detection, another Artificial Neural Network was used to identify nitrogen and yellow rust stressed wheat in addition to healthy winter wheat canopies. This model was able to accurately predict nitrogen stressed, yellow rust infected, and healthy canopies with 99.63%, 99.83%, and 97.27% accuracy, respectively. Silybum marianum, also known as milk thistle, is a common noxious weed that hinders crop growth. A model using an Artificial Neural Network used spectral bands of red and green as well as texture to identify milk thistle with 98.87% accuracy.</p><p>Another impressive task which machine learning has done a pretty good job at is identifying the classification and geographical origin of a rice sample using ensemble learning and random forest methods and does so with 93.83% accuracy.</p><p>This is an interesting article to look over because it gives somebody like me a lot of ideas of what machine learning and data science in general could be used for within the agricultural industry. It is clear that data science has taken over some industries such as automobile manufacturing and social media platforms/advertisements, however, it has yet to really make its way into agriculture. I understand that technology has made large leaps forward, especially when comparing a harvester (or combine) from, say, 50 years ago to a brand new one today. However, data science doesn’t seem like it has fully gotten its foot in the door. This is important because, while advancement in technology can make our lives easier, advancement in data science within an industry can increase our knowledge which can lead to even greater advancements than we might have previously thought.</p><p>This topic really interested me not only because of my past (education and stories from my dad) but also because of my future. After finishing the data science program at Flatiron, I think it would be especially rewarding to get a job in the agricultural or food industries because of their unbelievable importance. I also think that there is a lot of room to make a difference. For example, when you think about the distance that a lot of the food that you eat has to travel, it really makes you wonder if there isn’t a better way to be growing and moving what we see on our plate every night.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a88cc12ae2d4" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[A Sea of Tools in an Ocean of Data: Seaborn]]></title>
            <link>https://medium.com/@teigenolson/a-sea-of-tools-in-an-ocean-of-data-seaborn-e00e80ad1441?source=rss-c7e664d75f83------2</link>
            <guid isPermaLink="false">https://medium.com/p/e00e80ad1441</guid>
            <category><![CDATA[python]]></category>
            <category><![CDATA[seaborn]]></category>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[data-visualization]]></category>
            <dc:creator><![CDATA[Teigen Olson]]></dc:creator>
            <pubDate>Tue, 19 Oct 2021 03:22:37 GMT</pubDate>
            <atom:updated>2021-10-19T03:22:37.069Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/420/1*5AocIGagjtqBqF4WXRrErA.png" /></figure><p>Seaborn, the python digital visualization library and an extension of matplotlib, is an important tool in creating visuals to help explain your data. But to start, I think it would be best to touch on why data visualization is so important.</p><p>Visualizing data can be helpful to me when I take my first look at what I am working with. We as humans are especially good at noticing patterns, however, it is hard to do this when faced with thousands or millions of data records. Looking blankly at these records, wondering what you’re supposed to do with such a mess of data is where data visualization typically starts, and although this can be helpful in determining a goal or deciding first steps, I would argue that it is even more important elsewhere. Data scientists are employed in order to unpack and analyze huge amounts of data and formulate an actionable and (most usually) profitable finding from said data, therefore, it is important for them to use their findings to help persuade management to follow their recommendations. However, not everyone is well versed in statistics or data analysis and because of this, it is crucial to present your findings in a manner which is understandable by all — the clearer your results are from looking at a visual, the better.</p><p>This is where seaborn comes into play. Seaborn is great because it provides professional level visualizations in a somewhat easy to use format. Lets model this slight tutorial as if we were just handed the dataset. Here, we will be looking at the historical price and volume data of each day for Bitcoin. After we load the data and import the necessary libraries for working with the data (most commonly pandas, matplotlib, numpy, and seaborn) we can use the .corr() function to get an idea of which columns are more and less related (although this is not a direct part of Seaborn, I figured it would be important to include in this section regarding data exploration). First, however, we should identify our business problem as well as our target variable. In this case, we are interested in determining if there are any indicators which will give insight into the cryptocurrencies closing price of each day. From looking at the correlation table, we can see that the column which is most correlated with the closing price is the volume of bitcoin that was traded for that specific day in currency. The correlation coefficient for this pair was 0.344, indicating, weakly if I may, that there is a positive correlation between the two. Another way which we can <em>visualize</em> these relationships is by using the seaborn .pairplot() function. This will provides us with a number of graphs, each representing a unique pair of columns. From looking at this, we can see that each of the price categories (High, Low, and Close) are strongly related to each other, which we would expect.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*0_xjP10NDFhpFkpuB6jxsw.png" /></figure><p>Though, thanks to the handy .corr() function, we can determine which pairs were interested and refer back to the pairplot to take a closer look.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*wOSgDyLk4dH3WSPNBno4Hg.png" /></figure><p>In this case, we are interested in taking a closer look at the relation between the volume in currency and the closing price of the cryptocurrency. To do this, we will use the .relplot() function in seaborn.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ZXX-um1qkfPJsGsLlUrDYQ.png" /></figure><p>This is actually a bit discouraging, as we can see that the relationship is not very strong. We would definitely have to make some adjustments in our analysis to bring any fruitful recommendations to our superiors, as visuals don’t do their job if they don’t make their message clear. As you can see, the plot which was made is a scatter plot, however, if we want, we can indicate “kind=’line’” within the parentheses of the relplot function.</p><p>If you happen to be working with categorical variables, another helpful function is the catplot() function. This helps to get an understanding of where categorical variables lie in terms of values. These plots display where the bulk of the values lie and show the interquartile range. The other thing that these plots help to show are outliers, therefore, they can be important in determining how you might want to clean your data. Similarly to the relplot function, you can specify different kinds of catplots to use. The default is a box plot, however, there are also strip plots, bar plots, violin plots, and four others to choose from. My personal favorite from these is the violin plot, as it is a combination between the box plot and kde plot, so it creates a really good looking display of the distribution of categorical data.</p><p>The third important function that we can use to create visualizations in seaborn is the displot() function. This is used to show distributions of data, and, similarly to the other two functions, has different kinds as well. The kinds that can be utilized within the displot function are histograms, kde plot, ecdf plots and rugplots.</p><p>One parameter which can be altered in many seaborn plots is the ‘hue’ parameter. This is especially useful in categorical data analysis as, when used, it shades different categories different colors in order to differentiate between the two in the visual more easily.</p><p>Another thing we can do with the graphs is change the color of the background. We do this by passing sns.set_style(‘darkgrid’) prior to creating our graphs.</p><p>Clearly, there is a lot to learn in Seaborn, so if you haven’t already heard about it and are in the field of data visualization, I strongly recommend taking a look into the library.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e00e80ad1441" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[A Brief Look into Amazon Web Services]]></title>
            <link>https://medium.com/@teigenolson/a-brief-look-into-amazon-web-services-9f78a20881e1?source=rss-c7e664d75f83------2</link>
            <guid isPermaLink="false">https://medium.com/p/9f78a20881e1</guid>
            <category><![CDATA[google-cloud-platform]]></category>
            <category><![CDATA[amazon]]></category>
            <category><![CDATA[amazon-web-services]]></category>
            <category><![CDATA[microsoft-azure]]></category>
            <category><![CDATA[web-services]]></category>
            <dc:creator><![CDATA[Teigen Olson]]></dc:creator>
            <pubDate>Tue, 28 Sep 2021 03:53:25 GMT</pubDate>
            <atom:updated>2021-09-28T03:53:25.329Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*pKkBP8kcZOJ1fbk8Yfyh0Q.jpeg" /></figure><p>While I was researching the various data science libraries, I really wanted to choose one which I thought was different or particularly interesting. I investigated a few different libraries and as I did, I realized that I had heard and seen a lot about Amazon Web Services (AWS) but knew almost nothing about. Seeing that it was obviously an important tool for many businesses, I decided to take a closer look at AWS, what it is, and how it helps companies.</p><p>To get a general understanding of what AWS is, I started by watching a 10-minute tutorial video on Youtube. This is admittedly the second video I watched about it, the first video published by Amazon themselves was best summarized by one of the top comments ‘I watched the whole thing and learned nothing.’ The video was full of statements that sounded great but really did not provide a whole lot of information that would make sense to someone with little to no experience with the application thus far. From the second video, however, I found out that the service is used by about 80% of Fortune 500 companies. Not surprisingly, the Amazon offered service is not free, though one aspect which rather is surprising or different about the service is the way in which customers pay, hourly and region determined. Amazon also incentivizes companies to continue use of the service by reducing prices for longer term purchases.</p><h3><strong>Overview</strong></h3><p>I am going to give my best attempt at explaining some of the main aspects which AWS has to offer to companies, however, there are over 130 services offered through the Amazon subsidiary so this might brief in the grand scheme of things.</p><p>AWS is a cloud service which contains over 200 services. These services fall under 26 categories which range from Analytics to quantum technologies to virtual and augmented reality. Instead of having customers pay for the entire package which they offer, they are able to pick and choose the services they need most, paying per second or for longer periods of time to save money. Another way in which customers can ‘save’ more is by bundling services together.</p><p>Arguably the most commonly utilized AWS service is the Amazon Elastic Compute Cloud (Amazon EC2). EC2 is basically a way for customers to rent virtual computers to run their programs on. Users can choose their processor, storage amount, and even operating system. Costs range from $0.0058 per hour to as much as $4.992 per hour depending on the ‘instance’.</p><p>Another service which might be more familiar is Amazon’s Simple Storage Service (S3). This is similar to Apple’s iCloud, and is used to store big data. S3 is great because besides being incredibly secure, it stores your data redundantly in three separate data centers within your region in case something happens. The cloud service is priced in a pretty typical fashion, with customers paying by size of data being stored as an example, storing 500 GB on the standard S3 cloud would cost $0.022 per GB per month.</p><p>When you think of AWS, I think it is safer to ask the question, “What don’t they offer?” rather than what they do. Let’s just take a look of some more of the services they offer and what they do in short, plain English. To do this, I am taking some of the services within a selection of the 26 categories which seem like they would be used more often.</p><p><strong>Analytics</strong></p><p>Data Integration (AWS Glue)</p><p>Data queries using SQL (Athena)</p><p>Fintech analytics (Amazon FinSpace)</p><p><strong>Blockchain</strong></p><p>Ledger Database ( Amazon Quantum Ledger Database)</p><p>Scalable blockchain network (Amazon Managed Blockchain)</p><p><strong>Business Applications</strong></p><p>Alexa for business</p><p>An equivalent to zoom for the workplace (Amazon Chime)</p><p>Email (Amazon WorkMail)</p><p>Mobile and web-based app builder without programming (Amazon Honeycode)</p><p><strong>Cloud Financial Management</strong></p><p>Cost analysis (AWS Cost Explorer)</p><p>Cost and usage budgets (AWS Budgets)</p><p><strong>Compute</strong></p><p>Virtual servers (Amazon EC2)</p><p>Virtual private servers (Amazon Lightsail)</p><p>Web-application management (AWS Elastic Beanstalk)</p><p><strong>Customer Engagement</strong></p><p>Cloud based contact center (Amazon Connect)</p><p>User engagement personalization (Amazon Pinpoint)</p><p><strong>Developer Tools</strong></p><p>Shell environment based in browser (AWS CloudShell)</p><p>Universal tool for AWS service management (AWS Command Line Interface)</p><p>Application analysis and debugging (AWS X-Ray)</p><p><strong>Game Tech</strong></p><p>Game server hosting (Amazon GameLift)</p><p>3D game engine integrated with AWS and twitch (Amazon Lumberyard)</p><p><strong>Quantum Technologies</strong></p><p>Quantum computing exploration and experimentation (Amazon Braket)</p><p><strong>Robotics</strong></p><p>Robotics application developer (AWS RoboMaker)</p><p><strong>Satellite</strong></p><p>Fully managed ground station (AWS Ground Station)</p><p><strong>Virtual and Augmented Reality</strong></p><p>VR and AR application builder (Amazon Sumerian)</p><p>From looking at this incomplete list of the services that AWS can offer businesses, more than one of the services could be useful to companies. A big question, however, is whether the services being provided by amazon have any competitors.</p><h3><strong>Competitors</strong></h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*bSxXly-ird5_PlqsyFRR9g.png" /></figure><p>I was genuinely surprised to find out that there are multiple worthy opponents to the AWS giant. The first of which being Google Cloud Platform (GCP). Google cloud also offers hundreds of services to meet business needs including computing, storage, network, and privacy features. One of the most challenging aspects to compare is pricing due to the many variables involved. These variables include but are not limited to: virtual machine instance, storage amount required, purchase plans, support tier chosen, and data center locations. One helpful tool that I have found is pricing calculators for each of the services. Although these might not be exact, they would give users a good estimate regarding the cost of each of the services depending on how they plan to use them.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/318/1*tgbsM8uhlWLRpnAo8ZQDnA.png" /></figure><p>The other main competitor with AWS is Microsoft Azure. Although they might be a bit biased, Microsoft claims that AWS is 5 times more expensive. Microsoft works to keep costs down by helping customers save by utilizing existing Microsoft licenses owned by the customer. Another statistic which has confused me a bit is where Microsoft claims that 95% of Fortune 500 companies use Azure. This confuses me because Amazon claimed that 80% of Fortune 500 companies use AWS, so what I am taking this as is that companies prefer some services from one company and other services through their competitor. According to one article, Microsoft is seen as the safe global bet, because, while it might not offer the depth and breadth of services that AWS does, Microsoft’s pricing strategy is easier for businesses to understand and not find them in an unexpected amount of debt.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*_RLg-L8vdVrFLpAjzDzNng.png" /></figure><p>Another image which I thought might be useful in comparing competitors is the one listed above. It shows more than the services which i looked into above, and ranks them on where they stand in the industry, indicating whether the service leans more towards being an industry leader, visionary, challenger, or simply a niche being filled.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=9f78a20881e1" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>