<?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 Anterra Kennedy on Medium]]></title>
        <description><![CDATA[Stories by Anterra Kennedy on Medium]]></description>
        <link>https://medium.com/@anterrakennedy?source=rss-b01a152e1cc2------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*CmWOTGTUMNJ40mZFrTvnbA.jpeg</url>
            <title>Stories by Anterra Kennedy on Medium</title>
            <link>https://medium.com/@anterrakennedy?source=rss-b01a152e1cc2------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Thu, 19 Nov 2026 23:47:43 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@anterrakennedy/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[Hidden Gems: Finding the Best Secret Trails in America]]></title>
            <link>https://medium.com/data-science/hidden-gems-finding-the-best-secret-trails-in-america-d9203e8ad073?source=rss-b01a152e1cc2------2</link>
            <guid isPermaLink="false">https://medium.com/p/d9203e8ad073</guid>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[regression]]></category>
            <category><![CDATA[python]]></category>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[hiking]]></category>
            <dc:creator><![CDATA[Anterra Kennedy]]></dc:creator>
            <pubDate>Thu, 06 Aug 2020 03:03:39 GMT</pubDate>
            <atom:updated>2020-09-16T20:42:47.683Z</atom:updated>
            <content:encoded><![CDATA[<h4>With Data Science</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*tl96SD7jbazfWoiWsZk3-A.jpeg" /><figcaption>(photo credit: David Marcu via <a href="https://unsplash.com/s/photos/david-marcu">unsplash</a>)</figcaption></figure><p>There are plenty of reasons why one would want to find solitude in the wilderness, from the therapeutic effects of being immersed in nature, to not wanting to contribute to trail degradation and soil erosion on busier trails.</p><p>Now more than ever the reprieve of the outdoors is greatly needed. But in a post-COVID 19 world, where it can be practically impossible to maintain proper social distancing measures when passing hikers on a narrow trail, it is especially important to find less frequented trails to hike.</p><p>I set out on a mission to use data science and machine learning to find the best little-known trails in America.</p><h3>The Approach</h3><p>If you’re anything like me, before you go anywhere or buy anything, you’re going to read all the reviews.</p><p>When I first approached this project, I wanted to answer the question, “What makes a trail good?” That is, what combination of features and statistics about a trail would lead to it having a high overall rating?</p><p>What I pretty quickly found out though, is that across the 35,000 hiking trails I scraped and analyzed, basically all of them were rated “pretty good” — that is, with an average user rating of 4.2 out of 5 stars and standard deviation of less than 0.6, it was really hard to distinguish which trails were excellent, and which were just okay, from their 5-star rating alone.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/308/1*b_Aci9yW6nHrEiu9N9oz1g.png" /></figure><p>What there was huge variation in across all the trails though, was their popularity as represented by the total number of reviews each trail had. While the vast majority of trails had only 100 or so reviews, a select few had several thousand! What was making these trails so popular?</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/396/1*hubdDPP8mE0EMCja7q2XfQ.png" /></figure><p>I thus pivoted to try to predict not the rating of a trail, but instead determine, via a data-driven model, the relationship between the various features of a given trail and its popularity. In finding commonalities, I could then apply that model to unpopular trails, to find which ones check all the same boxes and are likely to be great, even though they haven’t been discovered yet.</p><h3>Methodology</h3><ol><li>) With Selenium and Beautiful Soup, web scrape to obtain trail data about 35,000 trails in the United States. This included information about the length of the hike, its elevation gain, its location, and a list of all of the natural features (such as waterfall, wild flowers, paving) the trail had.</li><li>) Clean this data and create a Pandas DataFrame. This included one-hot encoding dummy variables for all of categorical feature columns.</li><li>) Utilize the VADER Sentiment Analysis module to analyze the text reviews via simple Natural Language Processing for each trail and determine a mean composite score.</li><li>) Use linear regression modeling methodologies including Statsmodels OLS to determine the relationship between a trail’s features and its’ popularity.</li><li>) Perform feature engineering and regularization via LassoCV to remove multicollinearity amongst those features and optimize the model.</li><li>) Apply that model to trails that are described as “lightly trafficked”, to find trails which would be expected to be popular based on their combination of features, but just haven’t been discovered yet.</li></ol><h3>Findings</h3><p>A linear regression model was fit to the trail’s stats with the number of reviews (and hence, popularity) serving as the target variable. The model yielded a list of the most influential features on a trail on it being popular. These included there being a <strong>fee</strong>, having a <strong>high sentiment analysis score</strong>, it being <strong>rocky</strong>, and having a <strong>scramble</strong> and <strong>no shade</strong>, amongst others.</p><p>I interpret those important features like this:</p><ul><li><strong>A fee</strong>: If the most popular trails have a fee to use, this indicates they are likely located inside National Parks. As many National Parks are closed due to COVID, or may be very busy, it is even more important to find alternatives.</li><li><strong>Sentiment analysis score</strong>: Since all trails have roughly the same score out of 5 stars, its hard to gather a lot of reliable information about their quality from this rating alone. By using natural language processing to analyze the written text reviews themselves, I was able to gain an actual useful metric in determining how people actually feel about the trail. The higher the score (on a scale of -1=very negative to +1=very positive), the stronger people felt positively toward the trail, which was super useful in finding hidden gems.</li><li><strong>Rocky/scramble/no shade</strong>: What this says to me is that the very popular trails take place above tree line! It’s on those more difficult hikes with higher elevation gain that you encounter these features. And with higher elevation, you’ll likely get better views! As it turns out, people love these tougher trails.</li></ul><p>The R² of this model was optimized to 0.19. Though this isn’t a very high score, you can see below that this is because the relationship between trail features and popularity simply isn’t linear. The residuals plot below showing the difference between the predicted popularity values and actual values demonstrates this pretty clearly (if this were linearly dependent, residuals would all fall in a fairly horizontal bar around 0!) So what’s actually determining a trail’s popularity if not it having all the right features of a popular trail?</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*dGphSNl2-S4lrT0XrPJT3w.png" /></figure><p><strong>My key finding </strong>was that in browsing trails, users are shown the trails with the most reviews first and foremost, which leads to a form of recursive confirmation bias. If all trails have roughly the same rating, users will turn to the reviews to determine whether a trail is good, will choose to do one with a lot of reviews, hence feeding in to the loop of making the very few busiest trails even busier. Meanwhile, other similar trails may have plenty of opportunity but go neglected.</p><h3>So What Makes a Trail Popular?</h3><p>There are tens of thousands of hikes in the United States that can be searched online. Typically, trails with the most reviews get the most hikes, and hence even more reviews; while lesser known trails may be just a good, but are harder to find on the website, and hard to know for sure whether they’ll be a good trail if they have so few ratings.</p><p>So what makes a trail popular? Ultimately, <em>the search algorithm does.</em></p><p>It’s time we break out of that feedback loop, and find some amazing alternative hikes where we can avoid the crowds. But how will you know if a trail is going to be worth your time? Well, I used Machine Learning to do that work for you.</p><p>I fit the best model on a subset of trails which were designated as being “lightly trafficked”, and the R² for these trails was 0.08. This was actually encouraging, considering that these are specifically a selection of trails which <em>aren’t</em> popular, but according to this, given their features, should be.</p><p>A potential area of future work for this project could be fitting a polynomial features model instead of a linear one. Early exploration into this method yielded a promising R² improvement to 0.26, but did induce some feature collinearity by duplicating features, that would need to be feature engineered out. I’m looking forward to continuing this work once I have more machine learning tools at my disposal! But I’m absolutely thrilled to present you with this list of the best lesser-known trails in America as my very first end-to-end data science project.</p><h3>Hike The Trails</h3><p>Check out the Hidden Gems in your State below!</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1004/1*3kwOlBTt9s_BruhVohj1Lw.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*TErj3T_ePX9xYGqf7oujDA.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1022/1*8OZ12KqLDNPMPjbr6JlTaw.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*yEBtP_MaI4Tx5M-lg7zjnQ.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1018/1*ScMNezZKB9CfXlhS0NkhNw.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1020/1*uBupdJfIXpkVHTILE8uKIw.png" /></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=d9203e8ad073" width="1" height="1" alt=""><hr><p><a href="https://medium.com/data-science/hidden-gems-finding-the-best-secret-trails-in-america-d9203e8ad073">Hidden Gems: Finding the Best Secret Trails in America</a> was originally published in <a href="https://medium.com/data-science">TDS Archive</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Python Environments, Oh My!]]></title>
            <link>https://medium.com/atha-data-science/python-environments-oh-my-a4c5b00084f4?source=rss-b01a152e1cc2------2</link>
            <guid isPermaLink="false">https://medium.com/p/a4c5b00084f4</guid>
            <dc:creator><![CDATA[Anterra Kennedy]]></dc:creator>
            <pubDate>Mon, 20 Jul 2020 15:21:45 GMT</pubDate>
            <atom:updated>2020-07-20T15:21:45.282Z</atom:updated>
            <content:encoded><![CDATA[<h3>June 22nd, 2020</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/492/0*S-lv_l7uO88EwHb_.png" /></figure><p>(image credit: xkcd)</p><p>Okay okay, maybe its not quite that bad; but I did run into my first version of this craziness!</p><p>I ran into some brief difficulty trying to execute NumPy within PyCharm; I had up until this point not needed access to NumPy in my learning of introductory Python, and had been using virtual environments for each project I created. When trying to access NumPy from a virtual env, Pycharm insisted I did not have NumPy on my machine. I did in fact have it, installed as part of Anaconda.</p><p>I thought I could navigate the issue by switching my interpreter from a virtual environment to my system Python, which until recently had been 3.8. However, when I downloaded Anaconda, it also installed Python 3.7. Switching Pycharm to 3.8, unsurprisingly, still did not have any packages like NumPy available. Then I recalled that I had earlier made 3.7 my default Python per the Metis program’s recommendation. So, I switched Pycharm’s default interpreter to Python 3.7… and still no NumPy! I then tried what Pycharm had listed as my Anaconda environment Python as my interpreter, thinking surely NumPy must be in the same downloaded environment… to yet again no avail.</p><p>Ultimately, I looked up the location of my default python from terminal (it was ~opt/anaconda3/bin/python3.7), set that path as my default interpreter manually inside Pycharm (it was not offering this particular system Python as an option on its own), and then, finally, I could access the the NumPy modules I so desired. How convoluted!</p><p>This all comes from, like the image above suggests, using different package managers at different times — homebrew, pip, anaconda, etc. — to all install different things… I uninstalled Python 3.8 (both versions, as apparently I had 2 at different locations urs/local/bin/python3.8 and /Library/Frameworks/python.framework/Version/3.8/bin/python3 as well…) for simplicity’s sake, and now have 2 copies of 3.7 on my machine (the anaconda env listed above, and another local at /usr/local/bin/python3.7). I think this will keep things straight for now, and I *believe* these are the only versions of Python remaining on my machine, but, I see how easy it can be to get tangled.</p><p>Any more senior data scientists with recommendations on how to keep all this straight, or your advice on best practice, I’d love to hear from you!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a4c5b00084f4" width="1" height="1" alt=""><hr><p><a href="https://medium.com/atha-data-science/python-environments-oh-my-a4c5b00084f4">Python Environments, Oh My!</a> was originally published in <a href="https://medium.com/atha-data-science">Atha Data Science</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Foobar Update and Adventures in Optimization]]></title>
            <link>https://medium.com/atha-data-science/foobar-update-and-adventures-in-optimization-ba162fce9ed2?source=rss-b01a152e1cc2------2</link>
            <guid isPermaLink="false">https://medium.com/p/ba162fce9ed2</guid>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[python]]></category>
            <dc:creator><![CDATA[Anterra Kennedy]]></dc:creator>
            <pubDate>Tue, 14 Jul 2020 18:40:33 GMT</pubDate>
            <atom:updated>2020-07-14T19:19:47.563Z</atom:updated>
            <content:encoded><![CDATA[<h3>June 11th, 2020</h3><p>So I’m back, writing from the other side of my time inside Google’s foo.bar, and ready to reflect on what a transformative learning experience that was.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*dU1gKOspspc-X9TCGmMwxw.jpeg" /></figure><p>Ultimately, this go around, I made it through the first 2 (of 5) levels. And honestly, I have no hard feelings about that at all — I began teaching myself Python two months ago, and I was able to engage with and pass several of the hardest coding challenges I’ve ever encountered, <em>published by Google</em>, and had an absolute blast doing it. So let’s review!</p><p>Level 1 I described in my last post, but will elaborate on to say that my first (“brute force”) approach to the problem did not utilize list comprehension at first, but rather for-loops and iterative appends to lists. Indeed, I had only just been introduced to list comprehension and this challenge was my first real application of it. The problem itself had to do with generating and accessing a very long concatenated string of prime numbers. Version 1 was definitely slower, and the dramatic improvement of performance seen by implementing list comprehension was a really rewarding and tangible way of learning the benefit of this method.</p><p>In the second level, optimization would become integral. Level 2 consisted of 2 challenges, the first of which involved accepting as arguments numbers of any base from 1 to 9, performing specified operations on them until the output values ended up in a repeating loop, and determining how many repeating values in that loop the output was alternating between. I solved this problem in about 3 hours, and with great efficiency, and was quite proud of my solution and ability to pass all the test cases.</p><p>Level 2’s second challenge proved a great deal more difficult. It was an interesting and tricky take on creating a sort of inverted Floyd’s triangle, but incrementing by 1 along the diagonal instead of row by row — which causes all of the triangular numbers up to (I determined to be sufficient) two times value of the index you are trying to access to end up along the “x-axis”. Creating this triangle proved to be quite the mental challenge, and when it was said and done I had done it with <strong>4 completely different approaches.</strong></p><p>The key difficulty with the problem was that the challenge wanted me to access which number was located at certain “x” and “y” coordinates, meaning find what number when appending incrementally larger numbers on the diagonal ended up in a certain row and column. Essentially, it wanted a structure which could be indexed in 2 directions, with indices of up to 100,000,000 each. And in all of my iterative designs, I had it working! Except, it was each time failing just 2 hidden test cases, which I imagine must have been very large numbers, since every version of my program would time out with very large inputs. Why you ask? Well, that all has to do of course with optimization…</p><p>My various designs included: first creating a Floyd triangle comprised of lists of numbers, then indexing those lists in such a way to build the desired triangle by rearranging the original triangle. Appending values to a set of lists and then appending to a <em>second </em>set of lists, definitely rough on the computational time and power. But, this brute force method was at least passing all but those 2 test cases, so from this framework I went about optimizing.</p><p>Other designs to try to mitigate the time-out-inducing processing complexity included:</p><ul><li>First appending to a dictionary with a for loop and then using dictionary comprehension to first make the Floyd’s and then the manipulated Floyd’s triangle, thinking that the use of both comprehension and dictionaries (must faster than lists since then can be accessed by keys instead of indexed!) would improve computation time</li><li>Using list comprehension followed by dictionary comprehension to do the same</li><li>Using only dictionaries to make the two consecutive triangles.</li><li>Trying to implement a generator to remember where in the first triangle creation the program was, to create the second from that place directly, without having to save all the numbers to memory (unsuccessful).</li><li>Using a double ended deque with a fixed length as a sliding window to try to do the same (unsuccessful).</li><li>Using numpy arrays (but I don’t really know anything about these yet and couldn’t get them to work).</li><li>Then finally shifting approaches <em>completely</em> and figuring out a mathematical formula to create the sequence of numbers in the inverted target triangle directly <em>without</em> first creating a Floyd triangle. The thinking was that I could hopefully cut the computation time in half (and it was a very tricky formula to derive!!) I created the inverted triangle directly, two ways: via two loops of appends to dictionaries, and another way with list comprehension.</li></ul><p><strong><em>They all failed </em></strong>the 2 hidden test cases.</p><p>I don’t even know for sure that it was because of computation time. I guess there’s no way of knowing what test cases I was failing as they were hidden, but none of the above would process fast enough to yield an answer for very large input numbers, so I assume it was still complexity issues.</p><p>But I’m not mad! I approached this problem <em>so</em> many different ways, and learned so much. In the future, I’m sure I’ll have more tricks up my sleeve that I can try, and get this to work.</p><p>Just after running out the clock on the 72 hour timer for the challenge, encountered this rather apt quote somewhere buried in a stack-exchange, on the subject of large data sets… “[they shouldn’t be a problem unless] naive use makes many small appends.” Well, call me naive — but only for now. I knew I needed to eliminate the repetitious appends which were sending my big(O) through the roof — I just didn’t know what to replace them with. And that’s exactly why I’m continuing to improve and build my skills.</p><p>I surprised myself with how very passionate about this I am — I was coding, continuously, about 12 to 15 hours a day, and doing so excitedly — which only affirms that I’m doing the right thing. And I proved to myself how many different ways I’m able to reshape and look at a problem, bringing in new creativity when I hit a wall. Last, I learned just how much new info I’m able to absorb and implement effectively on a pretty immediate basis. This was a huge moment of growth for me as a python developer.</p><p>I’ll be keeping the various entries on my github for reference: <a href="https://github.com/anterra/google-foobar/tree/master">https://github.com/anterra/google-foobar/tree/master</a> (don’t worry, Google randomly generates numerous versions for each applicant, so your ability to cheat off of these is very low. Plus, foobar is about the learning experience! Try it on your own 😉). I can’t wait to come back to the foo.bar someday in the future with more experience under my belt!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=ba162fce9ed2" width="1" height="1" alt=""><hr><p><a href="https://medium.com/atha-data-science/foobar-update-and-adventures-in-optimization-ba162fce9ed2">Foobar Update and Adventures in Optimization</a> was originally published in <a href="https://medium.com/atha-data-science">Atha Data Science</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Google FooBar]]></title>
            <link>https://medium.com/atha-data-science/google-foobar-626655339010?source=rss-b01a152e1cc2------2</link>
            <guid isPermaLink="false">https://medium.com/p/626655339010</guid>
            <dc:creator><![CDATA[Anterra Kennedy]]></dc:creator>
            <pubDate>Fri, 10 Jul 2020 18:32:13 GMT</pubDate>
            <atom:updated>2020-07-10T18:32:13.277Z</atom:updated>
            <content:encoded><![CDATA[<h3>June 6th, 2020</h3><p>There I was, studying for and completing my Metis program pre-requisite materials to be completed before the program begins later this month, when an innocent enough Google search to try to get a more in-depth understanding of list comprehensions triggered a Google browser easter egg I’d never heard of before — Google foo.bar. The usual page of search results tilted forward into the screen, revealing an enticing and mysterious invitation… of course I accepted, dead curious as I was. I’m loaded into a shell environment and given a sci-fi themed prompt asking for my help in completing a challenge, should I accept. At this point, not knowing anything about this, I go ahead and Google what this easter egg might be — and lo and behold its a quite reputable back-door invitation into interviewing to work at Google itself! Apparently, since of course they have your entire search history, they extend the invitation when your searches seem indicate proficiency or literacy in the types of skills they’re looking to hire for, and gets ‘triggered’ by certain search terms like my looking for list comprehension information.</p><p>There are 5 levels to the challenges, and I’ve just spent the last day completing the first. It involved prime numbers and — go figure — list comprehension (optionally, of course — it could surely have been completed with nested for loops and in fact that was my first approach as proof of concept). It was so fun!! Not terribly challenging but with several steps to work and think through. It served as a really nice refresher on math and number theory, showed me the advantages (processing power and runtime) of list comprehensions over for-loops in a really tangible way, and allowed me the chance to take a highly hands-on approach to more difficult programming and apply the concepts I’ve been learning to what feels more like the real world. At the very least it’s been a huge motivator! More updates to come…</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=626655339010" width="1" height="1" alt=""><hr><p><a href="https://medium.com/atha-data-science/google-foobar-626655339010">Google FooBar</a> was originally published in <a href="https://medium.com/atha-data-science">Atha Data Science</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[✨ Imminent Data Scientist ✨]]></title>
            <link>https://medium.com/atha-data-science/imminent-data-scientist-51e1444059cc?source=rss-b01a152e1cc2------2</link>
            <guid isPermaLink="false">https://medium.com/p/51e1444059cc</guid>
            <dc:creator><![CDATA[Anterra Kennedy]]></dc:creator>
            <pubDate>Fri, 10 Jul 2020 18:31:50 GMT</pubDate>
            <atom:updated>2020-07-10T18:31:50.633Z</atom:updated>
            <content:encoded><![CDATA[<h3>May 21st, 2020</h3><p>Long time no talk! I’ll explain why… it’s official, I’ve been accepted at Metis!</p><p>Been a bit pre-occupied with the application to update here, but it ultimately paid off, and now I’d like to reflect. The application process was three rounds — a traditional application where I wrote a couple of short essays about my background, my passion for data science, and my aptitude for the Metis program; a technical challenge which involved answering questions in statistics, probability, linear algebra, calculus, and data interpretation, solving 5 python coding challenges similar to those outlined in the last post, and writing up a project proposal for a data science passion project you might complete as your final project in the program; and finally a video call interview with a data scientist/Metis alumni, to discuss further my background, do some live statistics problem solving, and brainstorm further around the project proposal.</p><p>First of all, I think this whole process was amazing preparation for what technical job interviews will entail, so I’m grateful to have had the experience. Each portion of the application went really well and I felt confident in each.</p><p>I had spent the weeks prior to the application not only studying and practicing both python and mathematics (just needed some refreshing since university), but coming up with and fleshing out a project idea that I’m actually really excited about and proud of, and would potentially want to scale in to an actual application after the program. In short, my idea is to create a data-driven yoga sequencing application intended for yoga teachers. It would automatically generate unique and complete yoga sequences that would guarantee safe and effective transitions, leading up to and out of any peak posture chosen by the user. Preliminary research has indicated that a neural network and bi-directional LSTM modeling, along with a Markov chain, could be used to this end. I’m sure I’ll share more as I begin development on this program, but I’ll leave it at that for now.</p><p>Now that I’m officially in, everything feels so surreal, and tangible. I am actually on a fast-track toward the career and future that I’ve dreamed of. It’s been a real shift in my mental state and attitude — it feeling so real and immediate. I made a twitter and followed some of the lead data scientists at top companies, to start expanding my network, but more so to start being exposed to their opinions, best practices, and articles about innovation in the field they share. Already I’ve gained so much new information and gotten more of a feel for what other data scientists’ lives are like!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=51e1444059cc" width="1" height="1" alt=""><hr><p><a href="https://medium.com/atha-data-science/imminent-data-scientist-51e1444059cc">✨ Imminent Data Scientist ✨</a> was originally published in <a href="https://medium.com/atha-data-science">Atha Data Science</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Coding Challenges]]></title>
            <link>https://medium.com/atha-data-science/coding-challenges-67c635a81e28?source=rss-b01a152e1cc2------2</link>
            <guid isPermaLink="false">https://medium.com/p/67c635a81e28</guid>
            <dc:creator><![CDATA[Anterra Kennedy]]></dc:creator>
            <pubDate>Wed, 08 Jul 2020 05:30:50 GMT</pubDate>
            <atom:updated>2020-07-08T05:30:50.521Z</atom:updated>
            <content:encoded><![CDATA[<h3>May 3rd, 2020</h3><p>Yesterday and this morning I engaged in my first online coding challenges. And both challenging and fun they were!</p><p>As part of the Metis Program application process, I took an application readiness assessment, which involved a coding challenge hosted on HackerRank. It was my first online technical assessment of this kind, and my first time coding outside of PyCharm. While the actual problem to be solved did not strike me as terribly challenging, there are several reasons why I did not succeed on the assessment. Ultimately, I ran out of time — but primarily the time I did have was eaten up not so much by figuring out how to solve the problem, but in struggling with the unfamiliarity of the HackerRank environment. Finally I realized that they had pre-defined a __main__ method which used several test cases as input, and automatically loaded them into my functions and printed the results. As such, I did not have the ability to flow the input through different methods as I had expected (their method dictated what would happen to the input), nor view my output to debug how my program was working. By the time I figured all this out, and made changes to the __main__ method to complement my code, my time had expired.</p><p>Although I was disappointed to not have performed better on the pre-test, I didn’t see this as a setback, but a learning opportunity I was really grateful for. Undoubtedly in my job searches as a data scientist there will be countless technical challenges on a number of different platforms, each one with different parameters I’ll have to familiarize myself with. And I was thankful to have dealt with this only on the pre-test!</p><p>So, to best prepare myself for the actual technical assessment that would be coming up in the application, I decided to explore the HackerRank platform and learn how to interact with their coding environment, by trying out the Python certification test. And, actually — I passed! Not only did I get to reach a comfortable understanding of how their default __main__ method and test case entry worked, but I got to solve a challenging couple of problems, and did so successfully enough to be certified in Python! I didn’t expect to pass necessarily as I’ve only been practicing a couple of weeks, but just wanted to practice in the environment. So it was absolutely affirming and empowering to know that my practice had paid off and gained me a real, applicable and demonstrable skill. My approach this time was to actually create a program to answer the prompt in Pycharm first, in order to be able to monitor the output, see how it was actually functioning, and make incremental changes and improvements until it worked as intended against my own input. Then I copied that into the HackerRank code area and changed it so it would accept their input. And it was effective!</p><p>At this point, I feel prepared and excited for the actual application’s challenges.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=67c635a81e28" width="1" height="1" alt=""><hr><p><a href="https://medium.com/atha-data-science/coding-challenges-67c635a81e28">Coding Challenges</a> was originally published in <a href="https://medium.com/atha-data-science">Atha Data Science</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[#Why I’m Excited About Data Science!]]></title>
            <link>https://medium.com/atha-data-science/why-im-excited-about-data-science-6993a91f40a?source=rss-b01a152e1cc2------2</link>
            <guid isPermaLink="false">https://medium.com/p/6993a91f40a</guid>
            <dc:creator><![CDATA[Anterra Kennedy]]></dc:creator>
            <pubDate>Wed, 08 Jul 2020 05:29:00 GMT</pubDate>
            <atom:updated>2020-07-08T05:32:55.336Z</atom:updated>
            <content:encoded><![CDATA[<h3>Why I’m Excited About Data Science!</h3><h3>April 28th, 2020</h3><p>I wanted to take a moment to get down just why I’m so passionate and excited about the field of Data Science itself, how powerful I think it is to solve problems and create change, and why I think it will suit me so well.</p><p>But first, some other exciting things! First of all I turned 25 this week. It feels wonderful and surreal that in the past several weeks my life direction has manifested so clearly before me upon entering the latter half of my twenties; I feel trusting and confident in it, and so inspired!</p><p>Second, I bought a MacBook Pro! I’ve never used a Mac in my life, but already in just several weeks of beginning to learn Python on my (older) Windows laptop, I could see difficulty in integrating Python into a Windows environment, experienced a lack of command line capability and ease, and furthermore picked up on a clear preference toward Mac amongst the data science community. I was in need of a new laptop anyway, so I decided to make the switch so I can used to the new OS from the beginning of my new career’s journey, rather than getting any farther along in Windows and then needing to switch later on and hence being more out of my element and having more difficulty adapting. I love it! The learning curve wasn’t too steep, both the hardware and software are so gorgeous and the UI is really intuitive and pleasing!</p><p>Third, I discovered the Metis Data Science Bootcamp Program, and have made the decision to apply. The program is robust, intensive and comprehensive, and I know it would elevate my skills and prepare me for a career in data science in a way far beyond what I’d be able to do if entirely self-taught, and at a far more accelerated pace. Since I’ve already been so motivated and productive just self-teaching, I know I can channel this passion into the structure of the course. And furthermore because the course is so advanced, all of the self-led background research will prepare me for success in the program. Having identified this program as a goal has re-contextualized what I thought possible for myself on this path, reinvigorated my passion, to dream bigger.</p><p>Which brings us to data science and how excited I am to be entering this field.</p><p>I find data science to be so ubiquitous, and unbelievably powerful in its versatility. Data is indeed everywhere, and for all the times throughout various research projects I’ve been involved in I’ve felt there must be a more efficient and comprehensive method to look at the big picture and draw more meaningful conclusions — there is! And what’s more, data science is always evolving and adapting to the ever-changing needs of data solutions. I couldn’t be more excited to be entering a field where I can once obtain mastery, but also never stop learning. It also feels like the perfect middle ground between programming and science or engineering, where I can innovate and problem solve, yet my “laboratory” is my own laptop and program, the simplicity and power of which I love.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=6993a91f40a" width="1" height="1" alt=""><hr><p><a href="https://medium.com/atha-data-science/why-im-excited-about-data-science-6993a91f40a">#Why I’m Excited About Data Science!</a> was originally published in <a href="https://medium.com/atha-data-science">Atha Data Science</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[A Homework Assignment]]></title>
            <link>https://medium.com/atha-data-science/a-homework-assignment-cdce304fa8b?source=rss-b01a152e1cc2------2</link>
            <guid isPermaLink="false">https://medium.com/p/cdce304fa8b</guid>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[python]]></category>
            <dc:creator><![CDATA[Anterra Kennedy]]></dc:creator>
            <pubDate>Wed, 08 Jul 2020 05:27:59 GMT</pubDate>
            <atom:updated>2020-07-08T05:27:59.323Z</atom:updated>
            <content:encoded><![CDATA[<h3>April 23rd, 2020</h3><p>You never really know what you know (or don’t) until you have to do it yourself. That was demonstrated yesterday when a ‘homework’ assignment was posed by the textbook I’m working through, and I got to actually problem solve and write a program myself, instead of simply learn how various features of Python work through studying example programs.</p><p>The previous example had created a program which would check whether or not a user inputted string was a palindrome. The homework assignment was then to make the program more robust and accurate, by making it ignore punctuation, spaces, and capital letters, such that “Rise to vote, sir” for example, would be recognized as the palindrome that it is. The completed assignment can be viewed here: <a href="https://github.com/anterra/learning-python/blob/master/input_and_output/io_input_homework.py">https://github.com/anterra/learning-python/blob/master/input_and_output/io_input_homework.py</a></p><p>In designing how I would go about this, my first plan was to create a tuple of all the forbidden special characters, then iterate over that tuple to check if each one appeared in the input string, and if so delete it.</p><p>Then I planned to do the same with a tuple of capitalized letters, and if they appeared, replace them with lowercase versions of the same. I was trying to think of how to identify what sequence number the found capital letter held within the alphabet, so I could then replace with the same letter by calling the same position number in a lowercase alphabet tuple.</p><p>Both of these plans went through a few phases of trial and error before getting vastly simplified! First, I discovered the ‘string’ module, which includes built in ‘punctuation’ and ‘ascii_lowercase’/’ascii_uppercase’ strings. Excellent! Now I need not type out every possible punctuation mark etc.</p><p>So, I iterated over the punctuation string first and used:</p><pre>text.replace(character, “”)</pre><p>to delete any punctuation marks. However, when I then returned text, it was unchanged! I realized then that I must use :</p><pre>text = text.replace(character, “”)</pre><p>to reassign the variable as the changed version of itself, not just apply the change but not store it. I also included a separate line,</p><pre>text = text.replace(“ “, “”)</pre><p>to eliminate spaces, which are not included in string.punctuation.</p><p>Even further simplification came when I discovered Python’s .lower() method. No longer did I need to try to select which letter of the alphabet needed to be replaced with a lowercase version of itself, instead I could just convert the whole input string to lowercase in one simple line.</p><p>I ran the program with the example sentence, and it returned as a palindrome! It may be a very simple program, but it allowed me to get really familiar with string methods and applying a lot of the information learned so far, as well as the process of thinking through solving a problem with Python methodically and making optimization improvements, and was very satisfying!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=cdce304fa8b" width="1" height="1" alt=""><hr><p><a href="https://medium.com/atha-data-science/a-homework-assignment-cdce304fa8b">A Homework Assignment</a> was originally published in <a href="https://medium.com/atha-data-science">Atha Data Science</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Zip, Struggle, and Progress]]></title>
            <link>https://medium.com/atha-data-science/zip-struggle-and-progress-c87e9e56325c?source=rss-b01a152e1cc2------2</link>
            <guid isPermaLink="false">https://medium.com/p/c87e9e56325c</guid>
            <dc:creator><![CDATA[Anterra Kennedy]]></dc:creator>
            <pubDate>Tue, 07 Jul 2020 03:49:14 GMT</pubDate>
            <atom:updated>2020-07-07T03:49:14.999Z</atom:updated>
            <content:encoded><![CDATA[<h3>April 18th, 2020</h3><p>Just under a week now spent learning and working in Python, and with each passing day I’m more sure this is the path for me.</p><p>A lot of new skills have been acquired and developed in the last week, and I’ve slowly become more comfortable working inside Python. Its remarkable how much being inside a program is so like being inside a complicated math or physics problem, with an understanding of all the moving parts and ability to see how they connect and hold many parts at once toward a solution, your thoughts not taking the form of english but of pure logic. I’m thriving activating that part of my brain that hasn’t been truly stimulated in a couple of years.</p><p>I established a GitHub, and learned Git from the command line to commit and push version controlled programs up to my new repository with example and practice problems in it. Part of that involved learning to create a .gitignore file so as to avoid adding my entire venv to the repository. I also had the opportunity to relearn binary.</p><p>Working through A Byte of Python, I encountered my first challenge — to create a program which would create backups of your specified files and store them in a Zip file. The reason this posed difficulty was because Windows does not have a built in zip function. So, I turned to the third party 7-Zip. I set the path variables so ‘7z’ to zip files should have been executable from anywhere in the command line, but oddly it was only able to run when inside the 7-Zip folder. Some research online yielded this solution:</p><pre>PATH=%PATH%;C:\Program Files\7-Zip\<br>echo %PATH%<br>7z</pre><p>Which was successful in then allowing the 7z function to be called from anywhere (for instance, from inside the Python zipping program file). After lots of struggle, the program worked! It successfully created backup files and stored them in a zip file when executed from the command line. However, the program still couldn’t be successfully run inside PyCharm itself. Maybe in time I’ll develop enough expertise to get 7-Zip to be usable inside PyCharm — but even after adding it as a ‘tool’ it would not work.</p><p>The challenge prompt then asked that the program be rewritten instead using Python’s standard library zipfile function. This I carried out as follows:</p><pre>with ZipFile(target, “w”) as zip:<br>    for file in file_paths:<br>        zip.write(file)</pre><p>I had to learn through trial and error that the type of information the zip function needed was the actual paths to the files you want zipped, not just the files themselves. I ended up using os.walk to retrieve file paths to do so. Ultimately, I find using a built-in Python module to complete a task as opposed to relying on external system-specific functionality to be a much cleaner and more preferred solution to any problem. It allows cross-platform ability to run, by just downloading the Python program itself. Always seems best to be self-contained! At least for a simple program such as this.</p><p>What all of this demonstrated to me was some real world problem solving and debugging, and it was a joy!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c87e9e56325c" width="1" height="1" alt=""><hr><p><a href="https://medium.com/atha-data-science/zip-struggle-and-progress-c87e9e56325c">Zip, Struggle, and Progress</a> was originally published in <a href="https://medium.com/atha-data-science">Atha Data Science</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Does Ouroboros Have a Beginning? Getting Starting with Python]]></title>
            <link>https://medium.com/atha-data-science/does-ouroboros-have-a-beginning-getting-starting-with-python-b237ab4963cb?source=rss-b01a152e1cc2------2</link>
            <guid isPermaLink="false">https://medium.com/p/b237ab4963cb</guid>
            <dc:creator><![CDATA[Anterra Kennedy]]></dc:creator>
            <pubDate>Tue, 07 Jul 2020 03:44:52 GMT</pubDate>
            <atom:updated>2020-07-07T03:44:52.225Z</atom:updated>
            <content:encoded><![CDATA[<h3>April 13th, 2020</h3><p>So I’ve determined the direction I’m orienting my life → towards data science.</p><p>Now it’s up to deciding upon the steps to get there. It’s very different, after having spent a lifetime in academia, to now be completely self-directed. I could see how it would be very easy to get overwhelmed by the endless amount of information out there and not even start. But I am motivated by an undercurrent of passion and determination toward a clear vision, and so we begin.</p><p>I have identified the following skills which are fundamental to data science: programming, data visualization, data analysis, interacting with databases, and collection of data from various resources. To this end, the learning plan I’ve developed begins with developing proficiency in programming with Python, then learn data visualization with MatPlotLib, data analysis with Pandas, databases with SQL, and finally learn web scraping. I am open to changes in this plan as I gain new information about industry standards and best practice, but this is an outline from which to start my education.</p><p>Why Python?</p><p>Python is the industry standard for Data Science. In exploring preference between Python and R, I still favored Python due to its more robust nature as a full programming language, more versatility, and its gateway into AI and Machine Learning. R, while potentially superior in data visualization, seems preferred mostly within academia, and I want instead skills that can translate to a wide range of problems across any industry. Furthermore, Python is known to be highly beginner-friendly, due to its straight-forward syntax.</p><p>Next came choosing an IDE. I went with PyCharm after reading many reviews of all the available options. It is made by JetBrains whose other IDEs are preferred by programmers I know and came in highest rated across many articles for 2020, its accessible to beginners, is color coded for ease of readability, its has high aesthetic value (which I’ll be honest, I care about), is highly functional with debugging available etc. and is open source. Plus, its name reminded me of the charm quark (my favorite quark, by far.)</p><p>After installing both Python3 and PyCharm, I encountered my first problem solving challenge! Setting my PATH variables such that I could execute my new “hello world” module from the Command Line, and relearning relevant cmd commands to navigate to the relevant folder to execute the program. It’s not much, but it was a good refresher and was rewarding to make work.</p><p>I decided to begin my studies by working through the textbook “A Byte of Python” by Swaroop.</p><p>One thing I had to decide upon early was quotation convention for strings. The PEP8 doesn’t have a recommendation for best stylistic practice concerning this matter and they functionally do the exact same thing — basically pick one, and stick to it. I have chosen double quotes and I’ll explain why. First of all they look more like text and I find them more aesthetically pleasing. Also, I feel its far more likely for a string to contain a single stray quotation mark, for instance as an apostrophe in a contraction, than a double quotation — so using double quotes to contain strings eliminates potential difficulty with the existence of a stray quotation trying to end the string early.</p><p>So now I’m set up with my program, textbook, and conventions, and am eager to learn!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=b237ab4963cb" width="1" height="1" alt=""><hr><p><a href="https://medium.com/atha-data-science/does-ouroboros-have-a-beginning-getting-starting-with-python-b237ab4963cb">Does Ouroboros Have a Beginning? Getting Starting with Python</a> was originally published in <a href="https://medium.com/atha-data-science">Atha Data Science</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>