<?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[Arjun Kalburgi - Medium]]></title>
        <description><![CDATA[Issa ‘bout growin’ 🍃 - Medium]]></description>
        <link>https://medium.com/arjunkalburgi?source=rss----6c7fb8d61589---4</link>
        <image>
            <url>https://cdn-images-1.medium.com/proxy/1*TGH72Nnw24QL3iV9IOm4VA.png</url>
            <title>Arjun Kalburgi - Medium</title>
            <link>https://medium.com/arjunkalburgi?source=rss----6c7fb8d61589---4</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Thu, 19 Nov 2026 16:06:11 GMT</lastBuildDate>
        <atom:link href="https://medium.com/feed/arjunkalburgi" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Playing the Infinite Job Search Game]]></title>
            <link>https://medium.com/arjunkalburgi/playing-the-infinite-job-search-game-4e4b4b406425?source=rss----6c7fb8d61589---4</link>
            <guid isPermaLink="false">https://medium.com/p/4e4b4b406425</guid>
            <category><![CDATA[careers]]></category>
            <dc:creator><![CDATA[Arjun Kalburgi]]></dc:creator>
            <pubDate>Mon, 27 May 2019 13:37:28 GMT</pubDate>
            <atom:updated>2019-05-27T13:37:28.350Z</atom:updated>
            <content:encoded><![CDATA[<blockquote>Recently I got an email expressing concern for finding a job, in my reply I explain what I call ‘<strong>The Infinite Job Search Game’.</strong></blockquote><p>Recently I got an email expressing concern for finding a job. The sender is about to start a degree in Computer Science and was worried that Computer Engineering would have been the better choice for more job opportunities at the end of his degree.</p><p>After debunking the difference between the two disciplines, my advice to him was to play the <em>infinite job search game</em>, the best way of finding a job. I first learned about <a href="https://www.youtube.com/watch?v=0bFs6ZiynSU&amp;t=516s">the infinite game from Simon Sinek</a>, where he explains what the game is and how companies should play it.</p><p>While searching for a job after my degree, I realized that networking and job search is also an infinite game.</p><pre><strong>The Infinite Job Search Game</strong></pre><pre>1. Get involved with something</pre><pre>2. Ask a question about it </pre><pre>3. Find someone to answer it</pre><pre>4. Ask them the question via Twitter/Email/over coffee/etc. </pre><pre>5. Make them your mentor (ask them if they&#39;ll meet with you and if you can approach them for future questions)</pre><pre>6. Repeat <br>6a. Contact the previous mentors every so often for coffee/questions/life updates</pre><p><strong>Mentorship is key to this game.</strong></p><p>This brings purpose to your networking, because it’s about finding and learning from mentors instead of simply trying to meet someone. The more you learn, the more questions you have, the more you try to meet people that can answer those questions and help you along your journey.</p><p>These days, I try to email/text/contact people from my (now long) list of mentors at least once a week - I’ll actively seek out questions to ask them. Thanks to context, having a conversation with someone can be better than Googling it.</p><p>Companies care about finding people that are good people, not really about smart people. They don’t care about degree’s or grades, they care about hiring someone that will learn and is fun to work with. Getting to know people will prove to them that you’re willing to learn and are enjoyable, if they’re good mentors they’ll help steer you to opportunities they think you deserve.</p><p>Overall, I’m really excited for the future Computer Science graduate that emailed me. I wonder how in 4 years his knowledge of The Infinite Job Search Game will boost his learnings and opportunities.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*used5UdqOzSWNxql" /><figcaption>Wish I knew the Infinite Job Search Game back in 2013</figcaption></figure><h4>A quick note on job postings</h4><p>I actually don’t look at job postings at all. They’re useless, and they’re slow. I’d rather them be thinking about me in their meetings while they’re creating the role itself (this has actually happened multiple times). That happens by getting to know them, and with referrals.</p><figure><img alt="" src="https://cdn-images-1.medium.com/proxy/1*k3x4eNfGaG54H9skwyKeSA.png" /></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=4e4b4b406425" width="1" height="1" alt=""><hr><p><a href="https://medium.com/arjunkalburgi/playing-the-infinite-job-search-game-4e4b4b406425">Playing the Infinite Job Search Game</a> was originally published in <a href="https://medium.com/arjunkalburgi">Arjun Kalburgi</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Flailing for Flutter — trying something new]]></title>
            <link>https://medium.com/arjunkalburgi/flailing-for-flutter-trying-something-new-3d2de2b11bbe?source=rss----6c7fb8d61589---4</link>
            <guid isPermaLink="false">https://medium.com/p/3d2de2b11bbe</guid>
            <category><![CDATA[flutter]]></category>
            <category><![CDATA[data-visualization]]></category>
            <category><![CDATA[experiment]]></category>
            <dc:creator><![CDATA[Arjun Kalburgi]]></dc:creator>
            <pubDate>Wed, 15 May 2019 18:59:17 GMT</pubDate>
            <atom:updated>2019-05-15T18:59:17.106Z</atom:updated>
            <content:encoded><![CDATA[<h3>Flailing for Flutter — trying something new</h3><blockquote>Looking to create a new app and I wanted to experiment with Flutter. My two concerns of <a href="https://medium.com/@mx_tino/flutter-themes-9cebc0fecd1d">themeing</a> and <a href="https://medium.com/flutter-io/beautiful-animated-charts-for-flutter-164940780b8c">charts</a> were alieviated by two blog posts, but let’s try it out!</blockquote><h3>Installing Flutter</h3><p>The dreaded step is always installing an SDK just for experimenting… but here we go.</p><p>Installation Link: <a href="https://flutter.dev/docs/get-started/install">https://flutter.dev/docs/get-started/install</a></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*-Qz-Ht_Wt7XtfUiBO2tehg.png" /></figure><p>I needed to fix some installation issues which Flutter helped me out with. For a front-end developer, this was really helpful.</p><p>Unfortunately it also took forever, thanks to XCode.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*MuJ50QXKz3DI7FsuyFtWxQ.png" /></figure><p>After forever, we’re now ready to go! (I’m choosing to work in VS Code)</p><h3>Andriod Emulator and iOS Simulator</h3><p>Before I could get to coding I had to set up the Android and iOS devices on my machine. Fortunately I had ones set up for previous projects, I just needed to launch them like so:</p><pre>$ flutter emulators<br>$ flutter emulators --launch Pixel_2_XL_API_28<br>$ flutter emulators --launch iOS Simulator<br>$ flutter run -d all</pre><p>This runs the starter app on both!</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*NriZDJE1wntOJXn0xB8j-w.png" /></figure><h3>Starter App</h3><p>Now to start messing around! I’m using <a href="https://flutter.dev/docs/get-started/codelab">this tutorial</a> provided from the Flutter team as a start.</p><p>With the tutorial, I created a list (useful) and started learning about Flutter’s state (not really though).</p><p>I continued on to <a href="https://codelabs.developers.google.com/codelabs/first-flutter-app-pt2/#0">part 2</a>, which had me add interactive icons to list items to save them, navigate to a page to view saved list items, and begin theming the app.</p><h3>Theming</h3><p>Theming is one of the big reasons I wanted to try Flutter and it’s got a theming framework I wanted to try.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*SGb5vC4lmEvxolDwI2Lm5g.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*1IDC32PB1JC_3w4vOUC7lQ.png" /><figcaption>Times and Blue</figcaption></figure><p>First try at theming was promising, even just with changing the font and the primary colour. I feel like it’s generally stuck to the typical Material Design layout with the AppBar shadow and the darker background colour. Fortunately, I was quickly able to fix those, but I’m not certain I can make Flutter bend to <strong>all</strong> my UI wishes, would need more exploration.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*740y6H3WRm8JEZPbSM8tcQ.png" /></figure><h3>Graphs</h3><p>Equally important as theming is the ability to have good graphs. For this I’m looking at <a href="https://pub.dev/packages/charts_flutter">charts_flutter</a> recommended by the quick tutorial <a href="https://medium.com/flutter-io/beautiful-animated-charts-for-flutter-164940780b8c">blog post here</a>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*k3PI_0RZuvvD9CQRGYHfTQ.png" /></figure><p>The tutorial is to get a bar chart up but I need a line chart with specific properties, which proved to be much more difficult thanks to some of the code from the bar chart tutorial messing me up. I had to redo the bar chart demo a couple of times and add in <a href="https://google.github.io/charts/flutter/example/time_series_charts/simple">this Line Chart demo</a> carefully.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*EJS1ev59fgo2DpKWB6vnTg.png" /></figure><p>I don’t care for my chart being interactive, so I dropped that extra fluff from the tutorial. What I need is to make the graph moveable and zoomable along the yaxis, which I don’t think <em>flutter_charts </em>will be able to do.</p><p>But I tried anyway, starting with a bunch of different tutorials from <a href="https://google.github.io/charts/flutter/gallery.html">here</a>. I actually ran into a lot of errors while trying to set all these up and while some of them were clearly due to my unfamiliarity with Flutter, some of them were very: 🤨. That’s a drawback.</p><p>I ended up finding the <a href="https://google.github.io/charts/flutter/example/axes/numeric_initial_viewport">Numeric Initial Viewport Axes Example</a> which did the scrolling y-axis for me and just needed to combine it with the <a href="https://google.github.io/charts/flutter/example/behaviors/selection_callback_example.html">Selection Callback Interactions Example</a> (which features a TimeSeriesChart, which I need). This actually proved to be quite difficult.</p><p>In the end I wasn’t able to get it to work seemingly due to a <a href="https://github.com/google/charts/issues/134">bug that hasn’t been fixed</a>. This is how it would’ve worked though, by setting up a viewport and allowing PanAndZoomBehavior.:</p><pre>new charts.TimeSeriesChart(<br>  widget.seriesList,<br>  animate: widget.animate,<br>  domainAxis: new charts.DateTimeAxisSpec(<br>     viewport: new charts.DateTimeExtents(<br>          start: new DateTime(2017, 10, 3), <br>          end: new DateTime(2017, 10, 10))),<br>  behaviors: [new charts.PanAndZoomBehavior()],<br>  selectionModels: [new charts.SelectionModelConfig(<br>     type: charts.SelectionModelType.info,<br>     changedListener: _onSelectionChanged,<br>  )],<br>)</pre><p>Turns out Flutter isn’t capable of the graphing I need :(</p><h3>Conclusion</h3><p>In conclusion, Flutter seems like a pretty cool way to make apps but unfortunately it’s chart library isn’t functional and customizable enough.</p><p>Looks like I’m sticking to Ionic Angular with plot.ly :)</p><h4>Follow me</h4><p>For more of my coding, designing, and problem solving journey, follow me on <a href="https://twitter.com/askalburgi">Twitter</a> and <a href="https://medium.com/arjunkalburgi">Medium</a>. I like to post about new things I’m doing and am interested in. :)</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=3d2de2b11bbe" width="1" height="1" alt=""><hr><p><a href="https://medium.com/arjunkalburgi/flailing-for-flutter-trying-something-new-3d2de2b11bbe">Flailing for Flutter — trying something new</a> was originally published in <a href="https://medium.com/arjunkalburgi">Arjun Kalburgi</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Learning CSS Grid for my Portfolio]]></title>
            <link>https://medium.com/arjunkalburgi/learning-css-grid-for-my-portfolio-7140f80ca936?source=rss----6c7fb8d61589---4</link>
            <guid isPermaLink="false">https://medium.com/p/7140f80ca936</guid>
            <category><![CDATA[css]]></category>
            <category><![CDATA[web-development]]></category>
            <category><![CDATA[css-grid]]></category>
            <dc:creator><![CDATA[Arjun Kalburgi]]></dc:creator>
            <pubDate>Thu, 25 Apr 2019 14:51:02 GMT</pubDate>
            <atom:updated>2019-04-25T14:51:02.322Z</atom:updated>
            <content:encoded><![CDATA[<blockquote>Been seeing a lot about CSS Grid on Twitter, decided to fix some smelly HTML/CSS on my portfolio using it!</blockquote><p>It seems like I’ve settled into a learning process. As with <a href="https://medium.com/arjunkalburgi/viewing-vue-with-theradlist-8851502ab709">learning Vue.js</a> and <a href="https://medium.com/arjunkalburgi/arty-notes-sketch-noting-my-thoughts-e29f142a9eb4">learning sketch-noting</a>, I started with content.</p><p>In this case it was this <a href="https://css-tricks.com/snippets/css/complete-guide-grid/">article</a> on CSS Tricks, it went over the basics of CSS Grid and how to use it.</p><pre>.container {<br>  display: grid | inline-grid;<br>  grid-template-columns: 40px 50px auto 50px 40px;<br>  grid-template-rows: 25% 100px auto;<br>}</pre><pre>.item {<br>  grid-column-start: &lt;number&gt;;<br>  grid-column-end: &lt;number&gt;;<br>  grid-row-start: &lt;number&gt;;<br>  grid-row-end: &lt;number&gt;; <br>}</pre><p>With this much alone I was able to begin using the Grid for what I needed. I didn’t need to go over grid-template-areas, grid gaps and item alignments. Although they would be really useful.</p><p>Instead I needed this CodePen found in the article:</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fcodepen.io%2Fchriscoyier%2Fembed%2Fpreview%2FxBmYJN%3Fheight%3D600%26slug-hash%3DxBmYJN%26default-tabs%3Dhtml%2Cresult%26host%3Dhttps%3A%2F%2Fcodepen.io&amp;url=https%3A%2F%2Fcodepen.io%2Fchriscoyier%2Fpen%2Fecff0af160b3fd32cf67841b1eb6cfc3&amp;image=https%3A%2F%2Fscreenshot.codepen.io%2F3.xBmYJN.small.09fcb69a-82f5-47d3-8d2b-d5b29349fbf2.png&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=codepen" width="800" height="600" frameborder="0" scrolling="no"><a href="https://medium.com/media/47de988ff43fa6c4c81f9d82ec3ed40e/href">https://medium.com/media/47de988ff43fa6c4c81f9d82ec3ed40e/href</a></iframe><p>I began working in my own CodePen and didn’t define my grid properly so I went to an example to see how they had done it.</p><p>Once I figured it out I went on to finish recreating the layout. That became v1, which you can see here:</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fcodepen.io%2Faskalburgi%2Fembed%2Fpreview%2FROByaj%3Fheight%3D600%26slug-hash%3DROByaj%26default-tabs%3Dhtml%2Cresult%26host%3Dhttps%3A%2F%2Fcodepen.io&amp;url=https%3A%2F%2Fcodepen.io%2Faskalburgi%2Fpen%2FROByaj%3Feditors%3D1100&amp;image=https%3A%2F%2Fscreenshot.codepen.io%2F2549184.ROByaj.small.d7ac0ab1-019a-4649-9634-afe0ba092f71.png&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=codepen" width="800" height="600" frameborder="0" scrolling="no"><a href="https://medium.com/media/97732bd2f2b8fd842172335dbad6c72b/href">https://medium.com/media/97732bd2f2b8fd842172335dbad6c72b/href</a></iframe><p>Next I needed to animate the images like on my Portfolio. I copied over the Javascript and attached the necessary Javascript files (jQuery and <a href="https://tgomilar.github.io/paroller.js/">Paroller</a>) to make the parallax effect.</p><p>This became v2 which is here:</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fcodepen.io%2Faskalburgi%2Fembed%2Fpreview%2FmgjjKR%3Fheight%3D600%26slug-hash%3DmgjjKR%26default-tabs%3Dcss%2Cresult%26host%3Dhttps%3A%2F%2Fcodepen.io&amp;url=https%3A%2F%2Fcodepen.io%2Faskalburgi%2Fpen%2FmgjjKR&amp;image=https%3A%2F%2Fs3-us-west-2.amazonaws.com%2Fm.cdpn.io%2Fscreenshot-coming-soon-small.png&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=codepen" width="800" height="600" frameborder="0" scrolling="no"><a href="https://medium.com/media/b6463de12d00d3647bf41e755a2966ef/href">https://medium.com/media/b6463de12d00d3647bf41e755a2966ef/href</a></iframe><p>Finally, I needed to make it responsive just like on my website. This is the real reason I wanted to update using CSS Grid. In the current version, I have additional HTML tags which I hide and show for mobile and desktop. I no longer wanted that extra HTML and CSS.</p><p>I began with removing the display: grid from the container and then followed through adjusting all the other elements.</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fcodepen.io%2Faskalburgi%2Fembed%2Fpreview%2FmgjGbM%3Fheight%3D600%26slug-hash%3DmgjGbM%26default-tabs%3Dcss%2Cresult%26host%3Dhttps%3A%2F%2Fcodepen.io&amp;url=https%3A%2F%2Fcodepen.io%2Faskalburgi%2Fpen%2FmgjGbM&amp;image=https%3A%2F%2Fscreenshot.codepen.io%2F2549184.mgjGbM.small.09b5a8a6-ff85-41bc-b890-2d8b253f8139.png&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=codepen" width="800" height="600" frameborder="0" scrolling="no"><a href="https://medium.com/media/e2054b8e1f67c7beb8aafc0bddbe06f5/href">https://medium.com/media/e2054b8e1f67c7beb8aafc0bddbe06f5/href</a></iframe><p>And just like that. We have our completely old website with a brand new grid layout system. And by golly was that easy and pumpkin pie. The grid system is awesome for sure.</p><p><a href="https://github.com/askalburgi/askalburgi.github.io/commit/00260e1192b41bc0e772f8c899e5ce8bae3d4ef2">This is the commit</a> where I make the move to the Grid System. In the HTML, I’m able to reduce 40 lines of code and in the CSS, I’m able to reduce ~11 lines.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ErpqfbqYAI0nxy1UXajRcQ.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*3g_xFvG3AO3u9fvzr7PGGQ.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*tT72BY7voH4t1u4L1uOIww.png" /></figure><p>Especially in the HTML, I was able to reduce a lot. Honestly it speaks more to the terrible HTML I had before than to how great the Grid System is, but either way, I was able to fix the issue that I had!</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*XnF2eYbJ32jpOj2pE6atHA.png" /><figcaption>Image for the caption lol</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/proxy/1*OoTeOBZ4tF09RWuXkXNihA.jpeg" /></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=7140f80ca936" width="1" height="1" alt=""><hr><p><a href="https://medium.com/arjunkalburgi/learning-css-grid-for-my-portfolio-7140f80ca936">Learning CSS Grid for my Portfolio</a> was originally published in <a href="https://medium.com/arjunkalburgi">Arjun Kalburgi</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Poem about Home]]></title>
            <link>https://medium.com/arjunkalburgi/poem-about-home-84b4c640e485?source=rss----6c7fb8d61589---4</link>
            <guid isPermaLink="false">https://medium.com/p/84b4c640e485</guid>
            <category><![CDATA[poetry]]></category>
            <dc:creator><![CDATA[Arjun Kalburgi]]></dc:creator>
            <pubDate>Wed, 24 Apr 2019 05:41:01 GMT</pubDate>
            <atom:updated>2019-04-24T05:41:00.990Z</atom:updated>
            <content:encoded><![CDATA[<blockquote>I wrote this poem on the plane from Edmonton to Toronto in 2019. At the time I was upset about the concept of ‘Home’ and what it means.</blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*3P3cs8-2V89x3DRn" /></figure><h4>What is Home?</h4><p>Home is where the heart is.<br>Even if you’re homeless there.</p><p>I struggled being homeless<br>But I had my friends to care.</p><p>I’m on my way home now<br>But I’d rather not go</p><p>Because home is where the heart is,<br>And my heart is back home.</p><p>🏠</p><blockquote>My blog is becoming more and more professional, hopefully I can always post my more artsy expressions.</blockquote><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=84b4c640e485" width="1" height="1" alt=""><hr><p><a href="https://medium.com/arjunkalburgi/poem-about-home-84b4c640e485">Poem about Home</a> was originally published in <a href="https://medium.com/arjunkalburgi">Arjun Kalburgi</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Arty notes: sketch-noting my thoughts]]></title>
            <link>https://medium.com/arjunkalburgi/arty-notes-sketch-noting-my-thoughts-e29f142a9eb4?source=rss----6c7fb8d61589---4</link>
            <guid isPermaLink="false">https://medium.com/p/e29f142a9eb4</guid>
            <category><![CDATA[design]]></category>
            <dc:creator><![CDATA[Arjun Kalburgi]]></dc:creator>
            <pubDate>Mon, 22 Apr 2019 14:16:00 GMT</pubDate>
            <atom:updated>2019-04-22T16:30:05.881Z</atom:updated>
            <content:encoded><![CDATA[<h3>Artsy notes: sketch-noting my thoughts</h3><blockquote>I’ve always been fascinated with sketch notes and bullet journalling. Recently I started trying my hand.</blockquote><p>Just like with learning Vue, my process to start sketch noting was to learn via video first.</p><p>I started with this YouTube playlist about sketch notes. They go over ideas of how you can create imagery for different phrases/concepts.</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2Fvideoseries%3Flist%3DPLCdugvAzY7lZMqV2VguMVPWEzW1aftQ1a&amp;url=http%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3D3Fi3kHDs0rg&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2F3Fi3kHDs0rg%2Fhqdefault.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=youtube" width="854" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/454c1c4983a15c039f85106db2b1ae44/href">https://medium.com/media/454c1c4983a15c039f85106db2b1ae44/href</a></iframe><p>Honestly I think there is probably better content out there to learn this, but Mike Rhode is the author of the Sketchnote Handbook so I watched his videos first. I started getting the concepts and applying them to my notes!</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*5eiLBfngln98WE4C" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*P45jNEcpZ88jT_6D" /></figure><p>Some are more art than others, while some have been more notes than art.</p><p>I want to continue learning more note/art techniques to keep my sketch-noting going! If you have any resources or advice, <a href="http://twitter.com/askalburgi">Tweet at me</a>!</p><figure><img alt="" src="https://cdn-images-1.medium.com/proxy/1*k3x4eNfGaG54H9skwyKeSA.png" /></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e29f142a9eb4" width="1" height="1" alt=""><hr><p><a href="https://medium.com/arjunkalburgi/arty-notes-sketch-noting-my-thoughts-e29f142a9eb4">Arty notes: sketch-noting my thoughts</a> was originally published in <a href="https://medium.com/arjunkalburgi">Arjun Kalburgi</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Better code thanks to Twitter]]></title>
            <link>https://medium.com/arjunkalburgi/better-code-thanks-to-twitter-8fad5d307557?source=rss----6c7fb8d61589---4</link>
            <guid isPermaLink="false">https://medium.com/p/8fad5d307557</guid>
            <category><![CDATA[animation]]></category>
            <category><![CDATA[codepen]]></category>
            <category><![CDATA[networking]]></category>
            <category><![CDATA[twitter]]></category>
            <dc:creator><![CDATA[Arjun Kalburgi]]></dc:creator>
            <pubDate>Sun, 21 Apr 2019 16:47:45 GMT</pubDate>
            <atom:updated>2019-04-21T16:47:45.317Z</atom:updated>
            <content:encoded><![CDATA[<blockquote>Recently I sent a tweet to a stranger that I follow on Twitter asking for his help. The result? The coding culture on the Internet is great.</blockquote><p>I’ve been really annoyed with the <em>leaves animation</em> on my portfolio.</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fcodepen.io%2Faskalburgi%2Fembed%2Fpreview%2FxQeLdx%3Fheight%3D600%26slug-hash%3DxQeLdx%26default-tabs%3Dhtml%2Cresult%26host%3Dhttps%3A%2F%2Fcodepen.io&amp;url=https%3A%2F%2Fcodepen.io%2Faskalburgi%2Fpen%2FxQeLdx&amp;image=https%3A%2F%2Fscreenshot.codepen.io%2F2549184.xQeLdx.small.f8b025fb-c938-496a-875c-a347d1a3f6d7.png&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=codepen" width="800" height="600" frameborder="0" scrolling="no"><a href="https://medium.com/media/71f1fb595b200e844e1e012b4086b6ee/href">https://medium.com/media/71f1fb595b200e844e1e012b4086b6ee/href</a></iframe><p>Though unnoticeable in use on my website, I hated how the leaves would bounce back and forth between two points but I didn’t really know how I could fix it (I’m new to animations).</p><p>I’ve been following <a href="https://medium.com/u/1aeed18fd27d">Varun Vachhar</a> on Twitter for some time and see his animation work all the time. Now that I moved my job search to Toronto, I decided to go out of my comfort zone and reach out to him about helping me on my animation. His response was more than I could’ve asked for!</p><h3>Arjun Kalburgi on Twitter</h3><p>@winkerVSbecks Hey Varun, sorry to creep you but I wanted your help on this animation. I want to further randomize the icon movement using GSAP. Currently they just go back &amp;amp; forth on a path, how do I make the path more complex so it looks more natural? https://t.co/cGNF4fdPtO</p><p>He ended up making a CodePen that forked mine, showing me that GSAP can re-run the animation after completion:</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fcodepen.io%2FwinkerVSbecks%2Fembed%2Fpreview%2FLvyVGM%3Fheight%3D600%26slug-hash%3DLvyVGM%26default-tabs%3Dhtml%2Cresult%26host%3Dhttps%3A%2F%2Fcodepen.io&amp;url=https%3A%2F%2Fcodepen.io%2FwinkerVSbecks%2Fpen%2FLvyVGM&amp;image=https%3A%2F%2Fscreenshot.codepen.io%2F149125.LvyVGM.small.a13df12b-7e0f-42b6-ad5e-29d6ff879b23.png&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=codepen" width="800" height="600" frameborder="0" scrolling="no"><a href="https://medium.com/media/5df7f4e52d414a35b3c8211fbed9abc4/href">https://medium.com/media/5df7f4e52d414a35b3c8211fbed9abc4/href</a></iframe><p>I made some tweaks to it and ended up with the following, which is now applied to my website! It’s a much more natural movement for the leaves! Thanks Varun!</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fcodepen.io%2Faskalburgi%2Fembed%2Fpreview%2FWWjmRM%3Fheight%3D600%26slug-hash%3DWWjmRM%26default-tabs%3Dhtml%2Cresult%26host%3Dhttps%3A%2F%2Fcodepen.io&amp;url=https%3A%2F%2Fcodepen.io%2Faskalburgi%2Fpen%2FWWjmRM&amp;image=https%3A%2F%2Fscreenshot.codepen.io%2F2549184.WWjmRM.small.c4986ed4-7f2d-43d5-94d4-92903ef97ab6.png&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=codepen" width="800" height="600" frameborder="0" scrolling="no"><a href="https://medium.com/media/c72af68b2fcfb272167acb41b3320f7f/href">https://medium.com/media/c72af68b2fcfb272167acb41b3320f7f/href</a></iframe><p>In that same thread, Varun also talked about using <a href="https://p5js.org/reference/#/p5/noise">p5 to animate noise</a>, something I’ve never even heard of. I asked him to meet for coffee so I could learn more and he told me all kinds of background information about animation frameworks that I never knew! I think it’ll really help me learn animations moving forward because up till now it’s really been throwing darts in the dark.</p><p>So I guess what I learned is that the process of randomly tweeting at people for code mentorship is pretty awesome! I’m definitely going to do it more.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=8fad5d307557" width="1" height="1" alt=""><hr><p><a href="https://medium.com/arjunkalburgi/better-code-thanks-to-twitter-8fad5d307557">Better code thanks to Twitter</a> was originally published in <a href="https://medium.com/arjunkalburgi">Arjun Kalburgi</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Stylist Assistant, updating stylists to do more]]></title>
            <link>https://medium.com/arjunkalburgi/stylist-assistant-updating-stylists-to-do-more-236bf74c50a7?source=rss----6c7fb8d61589---4</link>
            <guid isPermaLink="false">https://medium.com/p/236bf74c50a7</guid>
            <category><![CDATA[product]]></category>
            <category><![CDATA[capstone]]></category>
            <category><![CDATA[fashion]]></category>
            <category><![CDATA[computer-engineering]]></category>
            <category><![CDATA[ai]]></category>
            <dc:creator><![CDATA[Arjun Kalburgi]]></dc:creator>
            <pubDate>Wed, 17 Apr 2019 19:37:02 GMT</pubDate>
            <atom:updated>2019-04-17T19:37:02.271Z</atom:updated>
            <content:encoded><![CDATA[<blockquote>My final year capstone project helps personal stylists recommend daily outfits for their clients and shows potential for market disruption.</blockquote><h3>Introduction</h3><p>The Stylist Assistant is a companion for stylists and their clients. In the industry today, personal stylists work with their clients closely, often on a daily basis, helping them achieve the client’s desired personal style goal. This makes being a personal stylist a full-time job, and hiring a personal stylist very expensive. Hence, very few people have a personal stylist.</p><p>By reducing the stylist’s time commitment we can reduce the time and cost barriers for clients, thus increasing the demand in the industry.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*H2VpCIxNt9mqV1Jff74W1Q.jpeg" /></figure><h4>Collaborators</h4><p>Capstone projects at the University of Alberta’s Faculty of Engineering are usually in groups of 3–5. I worked on this project with <a href="https://www.linkedin.com/in/rui-wu-3812a2164/">Rui Wu</a> and <a href="https://www.linkedin.com/in/v%C3%ADtor-mendon%C3%A7a-221151b0/">Vitor Mendonca</a>.</p><h4>Objective</h4><p>The Stylist Assistant seeks to help stylists with their clients daily style using image detection to recognize clothing items, data science to match outfits and a mobile app to interface between the two parties.</p><p>The project’s potential to disrupt and grow the industry to a large degree makes it a potentially revolutionary technology.</p><blockquote><a href="https://docs.google.com/presentation/d/1U4SbZmzaiIkmf1hJq27UpBQrJmhd1VahhA49iJoCT1k/edit?usp=sharing">View presentation slides for Stylist Assistant here.</a></blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*wbElysDqE7oqwJZlm47dvQ.png" /></figure><h3>Image Detection</h3><h4>DeepFashion Dataset</h4><p>We found the <a href="https://liuziwei7.github.io/projects/DeepFashion.html">DeepFashion Project</a> in our research for existing solutions. It also studies the relationship between AI and fashion and so we used the project to help kickstart our solution.</p><p>DeepFashion was great for kickstarting us, but also for the <a href="http://mmlab.ie.cuhk.edu.hk/projects/DeepFashion.html">dataset</a> it provides. It had tagged data for all kinds of clothing features as shown below.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*DkyuFZrKj4CgmLzl.jpg" /></figure><p>Unfortunately, the tags for texture, fabric, shape, part and style were completely inaccurate and we could not use them in our solution. We made do with just category and continued with the project as a proof of concept.</p><h4>Neural Net</h4><p>We built our own neural net for the image tagging. Using the TensorFlow library as a solution that worked and satisfied the course requirements.</p><p>Our Neural Net was directly based off of <a href="http://www.image-net.org/challenges/LSVRC/2014/">Simonyan &amp; Zisserman’s entry to the ImageNet Competition 2014</a> — also the basis for DeepFashion.</p><h4>Alternatives Considered for Processing</h4><p>We wanted to explore creating a custom classifier with the IBM Watson API, but decided this route would not satisfy the workload and learning requirements for the course. In our quick test, using DeepFashion and the custom classifier worked well as the Watson API is assisted by its existing knowledge base.</p><blockquote><a href="https://github.com/askalburgi/stylist-assistant/tree/master/Project%20Source/neural-network">See the code for the Neural Net</a></blockquote><h3>Data Science</h3><p>The objective of the data science portion of the project is to combine outfits from the client’s wardrobe and to measure commonality between the stylist’s recommendation and the list of possible client’s outfits.</p><h4>Using DeepFashion Categories</h4><p>Through the list of all the possible clothing item categories in the DeepFashion dataset, we built 3 category trees in order to deduce similarity among items: Tops, Bottoms and Full-Body.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*xV3tu3251zPOZSr0_q0jzg.png" /></figure><p>Items in different trees have 0% similarity, items in different sections of the graph are 25% similar, items within the same section are 50% similar and items of the same category have 100% similarity.</p><h4>Similarity Matrix</h4><p>The clothing piece similarity allowed us to then make outfit similarity measurements which could then be used to find the best matching outfit from the stylist’s recommended outfit.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*SISeI0rwJZYMaDVVocYF3g.png" /><figcaption>The stylist’s recommended is on the right and the client’s wardrobe on the left.</figcaption></figure><p>Using outfit similarity we can retrieve the most similar outfit from the client’s wardrobe, resulting in an outfit (or outfits) for the client that satisfy the stylist’s recommendation.</p><blockquote><a href="https://github.com/askalburgi/stylist-assistant/tree/master/Project%20Source/recommender">See the code for the Recommendation API</a></blockquote><h3>Mobile App</h3><p>The Stylist Assistant mobile app is used by both stylists and clients, meaning a user can be both a stylist and a client using the same app (provided they have different email addresses). This design decision was made to help keep the project’s scope relatively narrow. Companies like Uber originally kept both types of users, rider and driver, on the same app, but then grew into two apps as both sides of the service expanded.</p><h4>UX Design</h4><p>Design patterns (visual layouts of the app views) were given less detailed thought, although some minor patterns which were more intuitive were implemented, such as tabs in the client side rather than a side menu. This follows the general design rule that discards the idea of keeping main features hidden in a side-menu navigation. The stylist side does feature a side-menu, but it is a much more acceptable design pattern for switching between clients. Similarly, the display of side-by-side cards was chosen over other list options for showing the wardrobe and recommended outfits of the client.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*djuOs4oNY9MJRgBT" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*mly3daHo0l2Lb1jX" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*NHoqSqItIBWuqQbh" /><figcaption>Left to Right: Home screen, Client’s recommended outfits, Stylist’s client details</figcaption></figure><h4>UI Design</h4><p><em>Note this project was done before taking the UI design course at RED Academy.</em></p><p>The visual aspects of the application, including the colours and fonts, were chosen to support emotional responses of reliability and elegance. The font, Eloquent Pro, and the colour black give an impression of sophistication. In colour theory, the use of navy is known to build feelings of trust and reliability.</p><p>These visual cues are present in the application in subtle ways, such as the tab-bar colour of the recommendations being navy to suggest that the recommendations are trustworthy. The rest of the application is a simple black and white to create a tone of elegance.</p><h4>Frontend</h4><p>We chose Ionic Framework for our app because it’s fast and reliable with simple UI tools that we could easily modify. After initializing views and navigation with default templating, everything was then redesigned to reflect our app mockups.</p><p>We worked incrementally like this for the rest of the app as well, first adding functionality, then styling it and fixing bugs.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*LlK-kfS_7MU-mu_X" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*91yXBox7Oj-F21AI" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*OG0DA3QKeXMPxC3N" /><figcaption>Left to Right: Home screen, Client’s recommended outfits, Stylist’s client details</figcaption></figure><h4>Backend</h4><p>Since Angular was being used in the frontend, we chose to use Firebase as our database in the backend. Firebase has a number of useful features and works great with Angular as both are Google-backed.</p><p>We used Firebase login to handle accounts and sign-in and to quickly show data within the app.</p><p>Firebase was great but there was a lot of compatibility issues because somehow we ended up with Angular v1 in our Ionic app working with a much higher version of Firebase.</p><blockquote><a href="https://github.com/askalburgi/stylist-assistant/tree/master/Project%20Source">See the code for the mobile app</a></blockquote><h3>Results</h3><p>We were ultimately unsuccessful in obtaining sufficient accuracy in our classification. I think if the DeepFashion dataset was what it promised to be, we would have had something really valuable on our hands. Neural nets really need consistent data to yield good results.</p><p>The project as it stands relies on stylists to help the app build fashionable outfits tailored to the client’s fashion goals. We believe that it would be possible for our system to work without the stylist’s input if there was a sufficiently large and accurate database. DeepFashion was just not that database.</p><h4>Key Learning Lesson</h4><p>Data is a great tool that can be leveraged to do great things, but clean and accurate data is really hard to come by and without it those great things aren’t.</p><p>Tell me what you think about the project via <a href="http://twitter.com/askalburgi">Twitter</a> or <a href="https://www.linkedin.com/in/arjun-kalburgi-82b25058/">LinkedIn</a>!</p><figure><img alt="" src="https://cdn-images-1.medium.com/proxy/1*OoTeOBZ4tF09RWuXkXNihA.jpeg" /></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=236bf74c50a7" width="1" height="1" alt=""><hr><p><a href="https://medium.com/arjunkalburgi/stylist-assistant-updating-stylists-to-do-more-236bf74c50a7">Stylist Assistant, updating stylists to do more</a> was originally published in <a href="https://medium.com/arjunkalburgi">Arjun Kalburgi</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Viewing Vue with TheRadList]]></title>
            <link>https://medium.com/arjunkalburgi/viewing-vue-with-theradlist-8851502ab709?source=rss----6c7fb8d61589---4</link>
            <guid isPermaLink="false">https://medium.com/p/8851502ab709</guid>
            <category><![CDATA[side-project]]></category>
            <category><![CDATA[javascript]]></category>
            <category><![CDATA[vuejs]]></category>
            <dc:creator><![CDATA[Arjun Kalburgi]]></dc:creator>
            <pubDate>Tue, 16 Apr 2019 21:12:50 GMT</pubDate>
            <atom:updated>2019-04-16T21:16:01.237Z</atom:updated>
            <content:encoded><![CDATA[<blockquote>Learning Vue.js by updating one of my more simpler projects, TheRadList, the Edmonton restaurant webapp that restores the excitement of finding a place to grub.</blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*Whk_KRBeoqPW-cYg" /></figure><h3>Objective</h3><p>TheRadList is a project I started in January of 2018. Every once in a while I’ll bring it back to update it or use it to learn something new.</p><p>Here’s the total history of TheRadList updates:</p><ul><li><a href="https://medium.com/arjunkalburgi/im-more-lazy-than-you-backends-with-google-spreadsheets-3eeac8d135c1">Update so that my friends can easily modify content</a></li><li><a href="https://github.com/theradlist/angular-radlist">Update to learn Angular 5</a></li><li><a href="https://medium.com/arjunkalburgi/theradlist-a-new-side-product-e07bf0d3fdd8">Initial launch of the project</a></li></ul><p>Now I’m using it to learn <a href="https://vuejs.org/">Vue.js</a>, one of today’s most popular web frameworks.</p><h3>Learning Vue</h3><p>Learning a new framework isn’t too difficult, since a lot of them take concepts from each other and most have great learning tools online.</p><p>The first thing I did was go to the <a href="https://vuejs.org/v2/guide/">Vue.js docs</a> to see what I could learn. There I found a video series from <a href="https://www.vuemastery.com/courses/intro-to-vue-js">Vue Mastery called “Intro to Vue.js”</a> which covers all the core principles of Vue.js and also builds a sample app along with you.</p><p>I didn’t bother with the sample app, because TheRadList is my sample app.</p><h3>Planning my Architecture</h3><p>After picking up all the core concepts and functionality of Vue.js, I looked at TheRadList’s HTML to see how I could refactor and update it with Vue.js.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*SVkBXGeYCdj6H6fw" /></figure><p>I know that the App Component needs to get <a href="https://vuejs.org/v2/guide/components.html#Emitting-a-Value-With-an-Event">emitted messages</a> from the Filter Component, which will then update the List Component via <a href="https://vuejs.org/v2/guide/components-props.html#One-Way-Data-Flow">props</a>.</p><p>If possible, I also want to make my filter functions capable of filtering through multiple properties. But that’s more of a stretch goal.</p><h3>Building out my App</h3><p>Refactoring should always be done incrementally and that’s essentially what this process is.</p><p>I began with creating components and slowly turning them into what I needed, using the documentation and the videos from the core principles course as a reference.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*h0h--08SQGxQUkJGg2uzXQ.png" /><figcaption>Git commits tracking my incremental progress</figcaption></figure><p>Slowly but surely I replaced all the functionality of the previous website. Starting with the component for the top hero section, then the component for the filter section and finally the rad list itself.</p><h3>Comments and Improvements</h3><p>It’s always great to document the things you learned and the things you can do better. Something that can nearly always be improved in the components and how they work together.</p><h4>More Components</h4><p>It’s possible more components might be helpful, especially the for UI. For example each filter in the Filter Component can be its own component, this would mean more component-oriented CSS, helpful for designers and developers both.</p><h4>Component Efficiency</h4><p>Inner-component structure is another way to improve your app. I updated how the filter-props (props that control the list filters) were passed into the List Component, yielding a much more readable approach that’s also good for performance. (the watch function is slow)</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*j5_TYqnW1RSfIW9R_7WiWw.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*WDLgd9EwIyG0jQELpGQV5Q.png" /><figcaption>Updating List Component to take in 1 prop for all the filters</figcaption></figure><h4>Arrow Function Quirkiness</h4><p>Took me quite a while to figure out why my watch function wasn’t working when trying to watch for when the user toggles the “Deep Dive” view.</p><pre>watch: {<br>     expand: (newVal, oldVal)=&gt;{ this.expanded = (newVal==&quot;large&quot;) }<br>}</pre><p>According to <a href="https://stackoverflow.com/a/40557674">StackOverflow</a>, it’s simply because arrow functions (()=&gt;{}) don’t take in bindings for this. In other words, this doesn’t work inside them! Just reverting back to regular function syntax fixed the problem. A good lesson.</p><pre>watch: {<br>     expand(newVal, oldVal) { this.expanded = (newVal==&quot;large&quot;) }<br>}</pre><h4>Single File Components</h4><p>Something I’d love to get to are <a href="https://vuejs.org/v2/guide/single-file-components.html">Vue’s Single File Components</a>, making Vue projects further like Angular and React with separation of concerns.</p><p>For a simple app like TheRadList, I didn’t see a point in adding this functionality, so instead I separated out my components into files. It helps with readability abstraction but isn’t nearly as beneficial as Single File Components themselves.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*9mbSlucUkIF34qBA" /></figure><h3>Results</h3><p>You can visit TheRadList at <a href="http://www.theradlist.com/">www.theradlist.com</a>, but you won’t see any difference (other than a few style bug fixes I found). But now you’ll know that the site uses Vue.js!</p><p>Tell me how it is via <a href="http://twitter.com/askalburgi">Twitter</a> or <a href="https://www.linkedin.com/in/arjun-kalburgi-82b25058/">LinkedIn</a>!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=8851502ab709" width="1" height="1" alt=""><hr><p><a href="https://medium.com/arjunkalburgi/viewing-vue-with-theradlist-8851502ab709">Viewing Vue with TheRadList</a> was originally published in <a href="https://medium.com/arjunkalburgi">Arjun Kalburgi</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[I’m more lazy than you — backends with Google Spreadsheets]]></title>
            <link>https://medium.com/arjunkalburgi/im-more-lazy-than-you-backends-with-google-spreadsheets-3eeac8d135c1?source=rss----6c7fb8d61589---4</link>
            <guid isPermaLink="false">https://medium.com/p/3eeac8d135c1</guid>
            <category><![CDATA[javascript]]></category>
            <dc:creator><![CDATA[Arjun Kalburgi]]></dc:creator>
            <pubDate>Sun, 14 Apr 2019 23:04:35 GMT</pubDate>
            <atom:updated>2019-04-17T05:47:00.554Z</atom:updated>
            <content:encoded><![CDATA[<h3>I’m more lazy than you — backends with Google Spreadsheets</h3><p>As most developers I’ve gone on a journey of how I can make websites faster and easier. I’ve experimented with WordPress, <a href="https://html5up.net/">HTML5UP</a>, Github Pages, and now I’ve got a new tool.</p><p>Say hello to <a href="https://sheety.co/">Sheety</a>, the shitty database API (as far as databases go) that’s also the best at being fast and easy.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ppV1HxVjysDSUgBcaL3HbQ.png" /></figure><p>Sheety was made by <a href="http://twitter.com/@phillipcaudell">Phillip Caudell</a> “because he hates making backends” and honestly it’s the greatest thing ever for front end developers like me.</p><p>Let’s look at an example. I’ve rebuilt my old project, <a href="https://medium.com/arjunkalburgi/theradlist-a-new-side-product-e07bf0d3fdd8"><strong>TheRadList</strong></a>, to use Sheety so that my friends can easily edit the restaurant content on the website.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*A8M_qmhSM5Lf3TZTr4LZjA.jpeg" /><figcaption>TheRadList with data from Sheety</figcaption></figure><h3>The Spreadsheet</h3><p>I’ve moved all the data from a JSON var in the JS of TheRadList (yes, that’s how far I went to avoid using a backend) to this Google Spreadsheet:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*VkNJX7cwCfKx3h7QCqwwGQ.png" /></figure><p>My friends have edit access to this spreadsheet and I made them a custom link to make it easier to find: <a href="https://theradlist.com/data">https://theradlist.com/data</a></p><h3>The API</h3><p>Putting the Google Docs Publish Link into Sheety’s API tool results in the following API link for me to use!</p><blockquote><a href="https://api.sheety.co/c1471c68-f293-46d9-ab4c-ca5e24e6e24e"><em>https://api.sheety.co/c1471c68-f293-46d9-ab4c-ca5e24e6e24e</em></a></blockquote><p>With all the data as objects in a list, it’s easy for me to read via JavaScript.</p><h3>The Javascript</h3><p>I copied the example on the Sheety website to load the data from the API.</p><pre>var template;  <br>$(document).ready(function () {<br>   $.getJSON(<br>      &#39;https://api.sheety.co/c1471c68-f293-46d9-ab4c-ca5e24e6e24e&#39;, <br>      function (data) {<br>         template = Handlebars.compile($(&#39;#item-template&#39;).html())<br>         $(&#39;#list-places&#39;).html(template(data))<br>      })<br>})</pre><p>It uses Handlebars to load the data into the specific #item-template I made and then into the div I need.</p><h3>Other Things</h3><h4>The Filters</h4><p>Unfortunately Sheety doesn’t have any endpoint logic, so I can’t retrieve specific objects or filter using the API. So instead I save the data into a variable and do simple list filter operations. Unfortunately none of the filter data is present at the moment so they all return an empty list.</p><h4>Nested Objects</h4><p>I did another project with Sheety where I needed to my data to have some nested objects and/or lists within the object. Sheety’s CSV style is obviously not suited for this. My solution was to create more Google Spreadsheets for the inner-data and use a primary_id to link all the data together. Then combine objects in JS onLoad. Who ever thought Database class would come in handy?</p><h4>Updating Data</h4><p>After the Spreadsheet is updated, Google/Sheety take some time for the changes to be reflected. Sometimes it’s instant but sometimes it takes a couple minutes. I don’t recommend using it for chat apps 😂</p><h3>The Results</h3><p>Visit <a href="https://medium.com/arjunkalburgi/theradlist-a-new-side-product-e07bf0d3fdd8"><strong>TheRadList</strong></a> today to see Sheety in action! Well it’s not that exciting, you won’t even see it, but it’s lazy and that’s what counts.</p><figure><img alt="" src="https://cdn-images-1.medium.com/proxy/1*OoTeOBZ4tF09RWuXkXNihA.jpeg" /></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=3eeac8d135c1" width="1" height="1" alt=""><hr><p><a href="https://medium.com/arjunkalburgi/im-more-lazy-than-you-backends-with-google-spreadsheets-3eeac8d135c1">I’m more lazy than you — backends with Google Spreadsheets</a> was originally published in <a href="https://medium.com/arjunkalburgi">Arjun Kalburgi</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Conversations with Wikipedia (Research Paper)]]></title>
            <link>https://medium.com/arjunkalburgi/conversations-with-wikipedia-research-paper-4b7717873087?source=rss----6c7fb8d61589---4</link>
            <guid isPermaLink="false">https://medium.com/p/4b7717873087</guid>
            <category><![CDATA[conversational-ai]]></category>
            <category><![CDATA[elearning]]></category>
            <category><![CDATA[chatbot-development]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <dc:creator><![CDATA[Arjun Kalburgi]]></dc:creator>
            <pubDate>Sat, 23 Mar 2019 20:55:25 GMT</pubDate>
            <atom:updated>2019-03-23T20:47:59.037Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*P4re_2qMolHcE0up.png" /></figure><p><strong>Abstract. </strong>Asking a peer to explain a concept is a simple example of learner-initiated conversation and an effective learning technique, but it is difficult to automate for many different users and topics. Research on conversational agents generally keeps to conversations of a more specific domain. In this paper, a learner-initiated conversational agent is introduced; it can converse with many users (one at a time) and is capable of discussing many topics. This poster presents this agent’s development with a specific focus on using conversational practices to help support informal learning experiences. The agent uses Wikipedia as a first step in the development of a flexible and mobile conversation-based learning experience.</p><blockquote>Arjun Kalburgi and Carrie Demmans Epp<br>University of Alberta, Edmonton AB T6G 2R3, Canada<br>{kalburgi, cdmmansepp}@ualberta.ca</blockquote><blockquote><strong>Keywords:</strong> E-learning, Conversational Agent, Wikipedia, Self-directed learning, Mobile learning, Informal learning.</blockquote><h4>1.<strong> Introduction and Motivation</strong></h4><p>Asking a peer to explain a concept is a simple example of a learner-initiated conversation that occurs naturally. This type of learning relies on a peer to have sufficient knowledge of the topic, which is not always the case. Automating the conversation means understanding the user’s goal and the user’s requested topic, for any user and any topic. The ultimate goal of this project is to create an automated educational experience that is useful and context-dependent, just like that of conversing with a peer.</p><p>Existing agents create and research experiences modelled after conversing with a tutor (Nye and Graesser and Hu, 2014). These agents are designed to help users with specific knowledge areas and under specific circumstances, but utilize concepts of dialog, interaction and natural language understanding (NLU) (<a href="http://psycnet.apa.org/record/2013-31546-001">Katz and Albacete, 2013</a>) that are applicable to any agent.</p><p>The project builds a prototype e-learning agent that attempts to explore the requirements of the desired experience. The prototype is built as a command line interface (CLI) and focuses specifically on the context and dialogue of the conversational experience. Content is supplied to the agent from Wikipedia so that a large number of topics can be covered.</p><h4>2. <strong>Existing Conversational E-Learning Agents</strong></h4><p>Conversational agents are still far from achieving human-like dialogue (<a href="http://aclweb.org/anthology/N/N13/N13-1098.pdf">Thomason and Litman</a>, 2013) despite researchers actively pursuing this goal. Dialog has been used to simulate the highly interactive nature of human tutoring (<a href="http://psycnet.apa.org/record/2013-31546-001">Katz and Albacete, 2013</a>), in specific domains where the conversational agent or tutoring system control most of the learning experience. However, it has been shown that there is no one size fits all language model to be used for such tutoring (<a href="https://ieeexplore.ieee.org/document/7756915/">Katz and Jordan and Albacete, 2016</a>). Conversational agents that help students learn by holding a conversation in natural language (Nye and Graesser and Hu, 2014) have been explored but conversations that are student-initiated have not. Existing agents ask students to explain topics, this project has the agent explaining topics from Wikipedia so the user can ask further questions.</p><h4>3.<strong> E-learning Dialog Development</strong></h4><p>A conversational prototype was built from an iterative, design-based research perspective. The prototype features three parts: prompting, briefly summarizing, and conversing with the user.</p><h4>3.1. <strong>Building Context using Wikipedia</strong></h4><p>The system needs to understand the goals of the user to have an appropriate conversation. These goals are the user’s context.</p><p>Wikipedia offered an approach to building this context with its large MediaWiki topology and consistently structured entries. Narrowing and filtering of information to satisfy the learner’s goals can be done by traversing the topology and scanning through the entries.</p><p>Both depth-first (i.e., choosing a Wikipedia page and exploring its content) and breadth-first (i.e., viewing the Wikipedia page in the topology, as it connects to other pages) explorations of the Wikipedia knowledge base were considered. To decide between these approaches, we conducted exploratory user evaluations.</p><p><strong>Initial Testing.</strong> The prompt, “what would you like to learn”, was used to start a discussion and observe the type of knowledge the discussion centered around. Centering around a specific topic signified depth-first interest, and centering around relations between topics signified breadth-first interest. A few college students were briefly asked the prompting question at a hackathon. The discussions displayed depth-first interest, thus the project continued using depth-first conversation.</p><h4>3.2. <strong>Creating the Agent with the Wikipedia API</strong></h4><p>Wikipedia API libraries were used to retrieve the information and metadata from a Wikipedia page for the agent to use. A search query helped identify a Wikipedia topic for the learner and a nested list implementation of a page’s sections and subsections was useful in leading the conversation to more specific information.</p><p>Although conversations at this stage were possible, the text retrieved from the sections is too large and lacked conversational flow. To improve the conversation, the text needed to be summarized.</p><h4>3.3. <strong>Creating Natural Conversation with the PyTeaser API</strong></h4><p>Interviews were conducted to better understand how the agent should create summaries and handle questions the learner may have afterwards. These 10 interviews were modelled after the existing prototype and were performed on college students.</p><p><strong>Extractive query focused summaries.</strong> Interviewers would summarize large text blocks using the visual and vocal feedback they received to ensure the conversation remained relevant to the interviewee. This model is called a query-focused summary.</p><p>The PyTeaser API was used to provide this type of summary in the prototype. The query was set to the subtopic title itself. This helped to create smaller, focused responses by the agent. However, like many other extractive summarization algorithms (<a href="http://wing.comp.nus.edu.sg/~antho/N/N10/N10-1133.pdf">Ceylan and Mihalcea and Ozertem and Lloret and Palomar, 2010</a>), PyTeaser’s results were often too long for them to be appropriate responses.</p><p><strong>Post-summary follow-up questions.</strong> During the interviews, follow-up questions were asked to help learners accomplish their learning goal.</p><p>This follow-up questioning was added to the agent after the summarization stage. Once the learner replies with his/her question, the agent responds by re-summarizing the topic, using keywords from the question as the query. Unfortunately the summarization again relied on PyTeaser, possibly bringing more confusion than help.</p><h4>3.4. <strong>Improving Conversation</strong></h4><p><strong>Using Concepts from Summarization. </strong>The agent can use NLU techniques to find keywords and concepts to help build a summary, similar to what the interviewer would do. The agent uses the IBM Watson NLU API (Watson) to retrieve concepts from the user’s response to help provide more data to the query focused summarizer.</p><p><strong>Nested Subtopic Summarization.</strong> Interviewers would include subsection headers when summarizing subtopics that had their own subsections. This helped interviewers build more context as to where the conversation should go. To replicate this within the agent, an additional summarization function was made using the same technique from Watson.</p><p><strong>Open-Ended Questioning Upon Exit.</strong> If interviewers felt that they had failed to deliver the information the participating learner was seeking, they would simply ask if that was the case. This sparked another conversational flow for the interviewer to try to accomplish the learner’s goal.</p><p>As this usually happened after the learner motioned for the interview to end, the agent asks similar questions at the end of the experience.</p><h4>4. <strong>Next Steps</strong></h4><p>There were many possible paths the prototype could take to further explore the potential for learner-initiated conversational agents.</p><p>The two APIs, PyTeaser and Watson, did not solve the original problem in most cases. Output could still be too long and lack conversational flow. Although the use of PyTeaser and Watson was serviceable for this stage of the prototype, both need to use context better to support real learning conversations.</p><p>Additionally, the project would benefit from being able to distinguish between conversations that meet or failed to meet the user’s learning goal. This type of metric would provide data that can be analyzed to improve the agent.</p><h4><strong>5. Conclusion</strong></h4><p>The prototype agent built shows the potential for learner-initiated conversations that are relevant and knowledgeable. Interviews helped establish the dialog and content of these conversations while NLU techniques helped were used to enable agent-learner conversation.</p><p>This exploration has the potential to be expanded on. The agent built is currently only able to discuss topics at a high level, but developing context-dependent summarization and more finely-tuned dialog would continue the exploration into deeper educational discussion.</p><h4><strong>References</strong></h4><ol><li>Hakan Ceylan, Rada Mihalcea, Umut Ozertem, Elena Lloret, and Manuel Palomar. 2010. Quantifying the limits and success of extractive summarization systems across domains. In <em>Human Language Technologies: The 2010 Annual Conference of the North American Chapter of the Association for Computational Linguistics</em>, pages 903–911, Los Angeles, California.</li><li>Katz, S., &amp; Albacete, P. L. (2013). A tutoring system that simulates the highly interactive nature of human tutoring. <em>Journal of Educational Psychology</em>, pages 105(4), 1126–1141.</li><li>J. Thomason and D. Litman, “Differences in User Responses to a Wizard-of-Oz versus Automated System,” <em>Proc. North American Chapter Assoc. Computational Linguistics: Human Language Technologies</em> (NAACL-HLT 13), 2013, pages. 796–801.</li><li>S. Katz, P. Jordan &amp; P. Albacete. Exploring How to Adaptively Apply Tutorial Dialogue Tactics. <em>In Proceedings of the 16th IEEE International Conference on Advanced Learning Technologies. </em>(ICALT2016), 2016.</li><li>Nye, B. D., Graesser, A. C., &amp; Hu, X. (2014). AutoTutor and Family: A Review of 17 Years of Natural Language Tutoring. <em>International Journal of Artificial Intelligence in Education</em>, <em>24</em>(4), 427–469. <a href="http://doi.org/10.1007/s40593-014-0029-5">http://doi.org/10.1007/s40593-014-0029-5</a></li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*3_pVQ9n_GCOr21L_hhVuJQ.png" /></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=4b7717873087" width="1" height="1" alt=""><hr><p><a href="https://medium.com/arjunkalburgi/conversations-with-wikipedia-research-paper-4b7717873087">Conversations with Wikipedia (Research Paper)</a> was originally published in <a href="https://medium.com/arjunkalburgi">Arjun Kalburgi</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>