<?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 Andy Zhang on Medium]]></title>
        <description><![CDATA[Stories by Andy Zhang on Medium]]></description>
        <link>https://medium.com/@andyzhang?source=rss-ebfbfc45654f------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*wr0KxpWq1Mnb0SMXg66oJQ.jpeg</url>
            <title>Stories by Andy Zhang on Medium</title>
            <link>https://medium.com/@andyzhang?source=rss-ebfbfc45654f------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Fri, 20 Nov 2026 13:19:00 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@andyzhang/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[How to build a website builder]]></title>
            <link>https://andyzhang.medium.com/how-to-build-a-website-builder-aead35ca5310?source=rss-ebfbfc45654f------2</link>
            <guid isPermaLink="false">https://medium.com/p/aead35ca5310</guid>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[software-engineering]]></category>
            <category><![CDATA[frontend]]></category>
            <category><![CDATA[product-development]]></category>
            <category><![CDATA[mvp-development]]></category>
            <dc:creator><![CDATA[Andy Zhang]]></dc:creator>
            <pubDate>Tue, 05 Sep 2023 17:23:28 GMT</pubDate>
            <atom:updated>2023-09-05T17:23:28.544Z</atom:updated>
            <content:encoded><![CDATA[<p>A website for building websites. Sounds easy, right?</p><p>It sure is if scoped right and if you have a great guide to break it down for you. Luckily, you’re reading a guide to build just that written by someone who built a website builder from scratch. By the end of this tutorial, you’ll be able to build a simple website builder (<a href="https://website-builder-demo.vercel.app/">demo</a>, <a href="https://github.com/andyzg/WebsiteBuilderDemo">code</a>). You’ll also learn how to think through decisions in the context of building a product.</p><h4>MVP Requirements</h4><p>Let’s break down the use cases of our website builder MVP:</p><ul><li>A user can view a read-only version of the website they built</li><li>A user can add new sections to their website</li><li>A user can edit the text content of their website</li><li>A user can view which mode they’re in (editing vs read-only)</li><li>A user can toggle between read-only and edit modes for their website</li></ul><p>There’s a million more ways to enrich our website builder but let’s stick with the basics. It should take 20 minutes and will require beginner knowledge in JavaScript, React and CSS.</p><h4>Out of scope</h4><p>The danger in product development is trying to build too much and not launching early enough. Let’s define what will be cut out of scope from our MVP so that you can finish this tutorial in 20 minutes:</p><ul><li>Styling the content</li><li>Server-side component</li><li>Database and data storage</li><li>Mobile-friendly</li></ul><p>The neat thing about this MVP is that it’ll be easy to deploy the MVP live as a client-side app. In other words, launch first then iterate later.</p><h3>1. Client-side design</h3><p>For the client-side app, keep it simple and organized. You’ll have a Website component that takes in two parameters: mode and websiteData. The mode will define whether the website is being read or edited. The websiteData will define the content of the website.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*rPy3G-TUw94ipvuviBAQ9g.png" /></figure><p>One observation is that websiteData and mode are stored at state in the Root component and passed to Website component instead of being stored as a state within Website component. This type of pattern is called a <a href="https://react.dev/learn/sharing-state-between-components#controlled-and-uncontrolled-components">controlled component</a>. While the uncontrolled component option is possible, the Website component’s only focus should be to render the website and the Root should have the ability to change the data such as the mode.</p><p>Below is some skeleton code to set up the client-side application. Note that the code below will be in JavaScript but the <a href="https://github.com/andyzg/WebsiteBuilderDemo">linked source code</a> is in TypeScript.</p><pre>// page.tsx<br>const Root = (props) =&gt; {<br>  // Store the two states `websiteData` and `mode`<br>  const [websiteData, setWebsiteData] = useState([{ text: &quot;Hello World&quot; }]);<br>  const [mode, setMode] = useState(&quot;edit&quot;);<br><br>  // Return the `Website` component with the data it needs to render<br>  return (<br>    &lt;Website websiteData={websiteData} mode={mode} /&gt;<br>  );<br>}<br><br>// website.tsx<br>const Website (props) =&gt; {<br>  // Props that&#39;ll determine how this component should be rendered<br>  const { mode, websiteData } = props<br><br>  // `sections` will contain a list of HTML elements with text<br>  const sections = []<br>  for (const section of websiteData) {<br>    // For each section in our data, add it to the list `sections`<br>    sections.push(<br>      &lt;div className=&quot;p-8&quot;&gt;<br>        {section.text}<br>      &lt;/div&gt;<br>    );<br>  }<br><br>  return sections<br>}</pre><p>The website is being stored as a list of objects where each object contains the text property containing the text of the section.</p><p>You can run this application on your computer by cloning this <a href="https://github.com/andyzg/WebsiteBuilderDemo">GitHub repository</a> and then going through the README instructions.</p><p>Demo: <a href="https://website-builder-demo.vercel.app/1-setup">https://website-builder-demo.vercel.app/1-setup</a></p><p>Source code: <a href="https://github.com/andyzg/WebsiteBuilderDemo/tree/main/src/app/1-setup">https://github.com/andyzg/WebsiteBuilderDemo/tree/main/src/app/1-setup</a></p><h3>2. Website editing UI</h3><p>If you want to make your website editable, two questions need answering:</p><ul><li>How does a user add a new section?</li><li>How does a user edit a section’s content?</li></ul><p>One of the constraints of a website builder is that the dimensions of the website should not be modified in the website builder. For example, if you add a button that said “Add section” between each section, the height of the entire website while in edit mode will be different than in read-only mode.</p><p>To overcome this constraint, add a button that hovers over the content itself without affecting the section’s dimensions. To build this, create a separate element that becomes visible upon hovering over the section.</p><h4>Adding an “Add section” button</h4><p>The secret sauce lies in the CSS and the position attribute. An element’s position is set to relative to define a new stack context. If that element contains a child with the position: absolute attribute, that child will be in the context of its parents’ stack. In other words, if you set left: 0 on the absolute element, it won’t hug the left side of the window but instead, the left boundary of the relative element.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*roOJJsSBhbjFiOZ0-rnxkw.png" /></figure><p>How does this concept apply to our use case? What you can do is make hovering buttons at the top and bottom of each section for adding new sections. Because they’re hovering, they won’t affect the dimensions of the website. This can be accomplished by abstracting all editing tools into a Toolbar component and then creating two buttons that hug the top and bottom of the section. The example below uses TailwindCSS to apply styling.</p><pre>function Website(props) {<br>  const { mode } = props;<br><br>  // Use `editing` variable to easily check if we&#39;re in edit mode<br>  const editing = mode === &quot;edit&quot;;<br>  // ...<br>  <br>  // ...<br>  // Each section now has a content element as well as a toolbar<br>  // One key change here is that the section&#39;s position is now set as relative<br>  sections.push(<br>    &lt;section className={&quot;relative group&quot;}&gt;<br>      &lt;Toolbar editing={editing} /&gt;<br>      &lt;div className=&quot;p-8&quot;&gt;<br>        {section.text}<br>      &lt;/div&gt;<br>    &lt;/section&gt;<br>  );<br>  // ...<br>}<br><br>function Toolbar(props) {<br>  const { editing } = props;<br>  <br>  // If we&#39;re not editing, then don&#39;t show any editing tools<br>  if (!editing) {<br>    return null;<br>  }<br>  <br>  // The toolbar is transparent by default (opacity-0)<br>  // If the cursor hovers over the section, then it becomes opaque (group-hover:opacity-100)<br>  return (<br>    &lt;div className={&quot;opacity-0 group-hover:opacity-100 transition-opacity&quot;}&gt;<br>      &lt;AddSection className=&quot;bottom-0&quot; /&gt;<br>      &lt;AddSection className=&quot;top-0&quot; /&gt;<br>    &lt;/div&gt;<br>  );<br>}<br><br>function AddSection(props) {<br>  const { className = &quot;&quot; } = props;<br><br>  // We&#39;re storing the classes in an array so that it&#39;s easier to read<br>  // by grouping the classes by function.<br>  const classes = [<br>    // Applying the absolute position attribute to the buttons<br>    // `absolute` sets the position as absolute<br>    // `left-1/2` adds `left: 50%`<br>    // `-translate-x-1/2` adds `transform: translate(-50%, 0)`<br>    `absolute left-1/2 -translate-x-1/2`, // Positioning<br>    `bg-blue-400 text-white rounded-sm cursor-pointer `, // Cosmetic styling<br>    `flex justify-center`, // Layout<br>    `w-6 h-6`, // Size<br>     className // Additional classes<br>  ]<br><br>  return (<br>    &lt;div className={classes.join(&quot; &quot;)}&gt;<br>      +<br>    &lt;/div&gt;<br>  );<br>}</pre><p>The most important line in the code above is this one:</p><pre>`absolute left-1/2 -translate-x-1/2`,</pre><p>These are classes provided by TailwindCSS that essentially create the below behavior and add the CSS documented on the right:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*aDV01CagmXUDk1YR6-CCJQ.png" /><figcaption>The TailwindCSS classes translate to the CSS properties on the right</figcaption></figure><p>If we open our app, we now get this:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1008/1*4dxduATrCSZrUm5SR9DF0A.png" /><figcaption>A section with two hovering “Add Section” buttons</figcaption></figure><p>Now we have a section that’s editable with buttons to add new sections above and below! Clicking on the buttons won’t do anything yet but they will in the next section. The hardest part of this guide is done now, time to take this to the finish line!</p><p>Demo: <a href="https://website-builder-demo.vercel.app/2-section">https://website-builder-demo.vercel.app/2-section</a></p><p>Source code: <a href="https://github.com/andyzg/WebsiteBuilderDemo/tree/main/src/app/2-section">https://github.com/andyzg/WebsiteBuilderDemo/tree/main/src/app/2-section</a></p><h3>3. Enabling website edits</h3><p>To enable website edits, two changes need to be added:</p><ul><li>When a user clicks on the “+” button, a new section needs to be added</li><li>When a user selects a section, the section’s text should become editable</li></ul><h3>Adding a section</h3><p>To enable this, attach an onClick event on both “+” buttons and modify the website’s structure. Do this by passing down a onAddSection() function that adds a new section to the website when a button is clicked.</p><p>There are many ways this can be approached: you can pass down the setWebsiteData function to the button and allow it to be responsible for updating the website data. You can also create abstractions so that each component needs to worry about only one responsibility:</p><ul><li>AddSection: Am I being clicked or not?</li><li>Toolbar: Where should the new section be added? Above or below?</li><li>Section: Pass the callbacks to the children.</li><li>Website: When a new section is added, how is the website’s data structure modified?</li></ul><p>Separating the concerns and ensuring that each component only has to worry about one thing will make it much easier to manage bugs in the future. For example, if there were many components including AddSection that were responsible for modifying the entire website and a bug was introduced, it would be hard to narrow down which component is causing the bug. Modularizing and minimizing the responsibilities of each component ensures that the code is as clean and organized as possible and improves testability.</p><h3>Editing section text</h3><p>One of the original constraints is to ensure that the website being edited has the same dimensions as the website that’s read-only. To ensure that the dimensions stay the same, use the same components if possible. In other words, if the read-only section is displayed with a div, let’s edit the content within the same div.</p><pre>function Section(props) {<br>  const { editing } = props;<br><br>  // ...<br>  return (<br>    // [section code]<br>    &lt;div contentEditable={editing} onInput={onSectionTextChange} /&gt;<br>    // [more section code]<br>  );<br>}</pre><h3>Debugging React behavior</h3><p>Upon testing this, you’ll run into a bug. Every time the content is edited, the cursor moves back to the beginning. To solve this, we do a quick search, decipher a highly upvoted answer into code, and test it again.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*57Uf-I_iE4btzGQkhIq0BQ.png" /></figure><p>When building MVPs, it’s easy to get caught up in the details of the work or of a bug but since the goal is to build an MVP in 20 minutes, a solution that solves the problem is all we need.</p><pre>function Section(props) {<br>  // ...<br>  const { onSectionTextChange } = props;<br>  const ref = useRef(null); // reference to the editable div<br><br>  // The below useEffect fixes the bug<br>  useEffect(() =&gt; {<br>    // Whenever the state is updated, update the div&#39;s text<br>    if (ref.current) {<br>      ref.current.textContent = sectionData.text;<br>    }<br>  // Run the useEffect whenever the section&#39;s text changes<br>  }, [sectionData.text]);<br><br>  return (   <br>    // [section code]<br>    // Note that we removed the child contents of the &lt;div&gt;. Instead,<br>    // we set the div content in the useEffect above<br>    &lt;div<br>      ref={ref}<br>      suppressContentEditableWarning={true}<br>      onInput={onSectionTextChange}<br>      contentEditable={editing}<br>      className=&quot;p-8&quot;<br>    /&gt;<br>    // [more section code]<br>  );<br>}<br><br>function Website(props) {<br>  // ...<br><br>  const onSectionTextChange = (sectionIndex, e) =&gt; {<br>    // Copy the immutable state. <br>    // JSON.parse(JSON.stringify(...)) is an easy native way to deep clone<br>    // an immutable object. In practice, use a library or be mindful of <br>    // how you update state. Read more here:<br>    // https://legacy.reactjs.org/docs/state-and-lifecycle.html#do-not-modify-state-directly<br>    const newWebsiteData = JSON.parse(JSON.stringify(websiteData));<br><br>    // Set the new value<br>    newWebsiteData[sectionIndex].text = e.target.innerText;<br>    setWebsiteData(newWebsiteData);<br>  }<br><br>  const onAddSection = (newSectionIndex) =&gt; {<br>    const newWebsiteData = [...websiteData];<br>    newWebsiteData.splice(newSectionIndex, 0, { text: &quot;Hello World&quot; });<br><br>    setWebsiteData(newWebsiteData);<br>  };<br><br>  for (let index = 0; index &lt; websiteData.length; index += 1) {<br>  sections.push(<br>    &lt;Section <br>      key={index} <br>      sectionData={sectionData} <br>      onSectionTextChange={onSectionTextChange.bind(this, index)} <br>      onAddSection={onAddSection.bind(this, index)} <br>      editing={editing} /&gt;<br>    );<br>  }<br>  // ...<br>}</pre><p>With that bug fixed, you now have a website editor that allows us to edit the website’s text and add new sections. You’re 90% there!</p><p>Demo: <a href="https://website-builder-demo.vercel.app/3-edit">https://website-builder-demo.vercel.app/3-edit</a></p><p>Source code: <a href="https://github.com/andyzg/WebsiteBuilderDemo/tree/main/src/app/3-edit">https://github.com/andyzg/WebsiteBuilderDemo/tree/main/src/app/3-edit</a></p><h3>4. Previewing our website</h3><p>The website is always in edit mode because there isn’t a way to toggle the mode between read-only and edit. For this last step, add a fixed element at the bottom of the window that will allow the website builder to toggle between edit and read-only mode.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*bA4yuSM3Qnodu86ueKNF8g.png" /></figure><p>The reason the fixed toggle bar is fixed at the bottom is because the other best options lead to a suboptimal user experience:</p><ul><li>If it has a fixed position at the top of the window, then it will block part of the first section because it is hovering over the website sections.</li><li>If it has absolute positioning at the top, then it will not be visible the moment a user scrolls a little bit.</li><li>If it’s positioned at the bottom of the window, the user is guaranteed to be able to view the top of the page and they can easily reference it at any scroll position.</li></ul><p>Add two elements to the new footer: text indicating whether the website is in edit or read-only mode and a button to toggle the modes.</p><pre>function Website(props) {<br>  const { mode, setMode } = props;<br><br>  // For now, we&#39;ll wrap the setMode function with onToggleMode.<br>  // This will simplify the logic in the footer to only check if the <br>  // button was clicked instead of figuring out what value to set next.<br>  const onToggleMode = () =&gt; {<br>    if (mode === &quot;edit&quot;) {<br>      setMode(&quot;read&quot;);<br>    } else {<br>      setMode(&quot;edit&quot;);<br>    }<br>  }<br><br>  // ...<br>  return (<br>    &lt;Fragment&gt;<br>      {sections}<br>      &lt;Footer onToggle={onToggleMode} mode={mode} /&gt;<br>    &lt;/Fragment&gt;<br>  );<br>}<br><br>function Footer(props) {<br>  const { mode, onToggle } = props;<br>  const classes = [<br>    &quot;fixed bottom-0 left-0 right-0&quot;, // positioning<br>    &quot;h-16 p-4&quot;, // sizing<br>    &quot;border-t-2 bg-white&quot;, // cosmetics<br>    &quot;flex justify-between items-center&quot;, // layout<br>  ]<br><br>  return (<br>    &lt;footer className={classes.join(&quot; &quot;)}&gt;<br>      {mode === &quot;edit&quot; ? (<br>        &lt;div&gt;Edit mode&lt;/div&gt;<br>      ) : null}<br>      {mode === &quot;read&quot; ? (<br>        &lt;div&gt;Read-only mode&lt;/div&gt;<br>      ) : null}<br>      &lt;button onClick={onToggle} className=&quot;bg-slate-600 text-white text-sm rounded-md p-2&quot;&gt;Toggle mode&lt;/button&gt;<br>    &lt;/footer&gt;<br>  );<br>}</pre><p>And with that, you’ve crossed the finish line! You’ve successfully built an MVP of a website builder! Your website builder is able to preview a website, modify a website’s text and sections, and toggle between the two modes.</p><p>Demo: <a href="https://website-builder-demo.vercel.app/4-mode">https://website-builder-demo.vercel.app/4-mode</a></p><p>Source code: <a href="https://github.com/andyzg/WebsiteBuilderDemo/tree/main/src/app/4-mode">https://github.com/andyzg/WebsiteBuilderDemo/tree/main/src/app/4-mode</a></p><h3>Next steps</h3><p>If you want to go the extra mile and try a challenge, here are some other features you can try building:</p><ul><li><strong>Color the section</strong>: Add a toolbar to each section enabling a user to set a background color and/or text color for the section.</li><li><strong>Saving the data</strong>: Once a user has built their website, they can ideally persist the data. Integrate the application with a database.</li><li><strong>Add columns</strong>: How can you have multiple columns of text in the same section? You’ll need to modify the data structure of the website as well as add additional UI elements to support adding a new column.</li><li><strong>Deleting sections</strong>: Add a button that removes a section</li><li><strong>Upgrade the editor</strong>: Instead of using a content editable div, you can integrate with a rich text editor library to allow richer formatting to the text.</li></ul><p>Thank you for making it this far! If you liked this guide, share it with a friend who’s also learning front-end development and leave a comment!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=aead35ca5310" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[A reflection on my time as an Uber Associate Product Manager]]></title>
            <link>https://andyzhang.medium.com/a-reflection-on-my-time-as-an-uber-associate-product-manager-ce9fccf1c66f?source=rss-ebfbfc45654f------2</link>
            <guid isPermaLink="false">https://medium.com/p/ce9fccf1c66f</guid>
            <category><![CDATA[product-management]]></category>
            <category><![CDATA[tech]]></category>
            <category><![CDATA[uber]]></category>
            <category><![CDATA[product]]></category>
            <category><![CDATA[reflections]]></category>
            <dc:creator><![CDATA[Andy Zhang]]></dc:creator>
            <pubDate>Tue, 15 Dec 2020 08:37:55 GMT</pubDate>
            <atom:updated>2020-12-15T08:39:37.323Z</atom:updated>
            <content:encoded><![CDATA[<p>Nothing could have prepared me for my time at Uber. Even after six internships, over a dozen hackathons, and an engineering degree, I wasn’t ready for my Uber ride. But to be fair, no one should ever feel ready to join Uber. It usually just happens and you get used to the discomfort of never feeling ready.</p><p>Fast forward to today, my Uber journey has officially come to an end. The last two and a half years carried me across three Uber teams, four cities across the globe, and a rollercoaster ride filled with learnings and memorable moments. As I officially end this crazy ride, I wanted to reflect on my time at Uber and share a glimpse of some of those moments and learnings.</p><h3>The APM Bootcamp Firehose</h3><p>Okay, maybe this APM bootcamp will prepare us to be product managers…</p><p>The APM boot camp squeezes decades of learnings and experiences into two weeks. We had dozens of tenure leaders across the company take time out of their day to talk to us. Breaks were rare, but so were dull moments. It reminded me of the days before final exams when students crammed months of knowledge during finals, hoping to become experts overnight.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*bg5xKIG7iI3G7u0kfG4ijQ.png" /><figcaption>Credit for graphic: <a href="https://www.doctorswhocreate.com/why-i-make-md-comics/">https://www.doctorswhocreate.com/why-i-make-md-comics/</a></figcaption></figure><p>Some sessions were supplemented with homework. Some were more entertaining and involved seeing who spent the most money on Ubers. Others were tedious but necessary and involved resolving a hundred support tickets.</p><p>We ended boot camp with a hackathon where everyone put their learnings to the test and presented in front of many product leaders. I worked on a solution to help facilitate the selection of a restaurant for a group order. Users would be able to join a group order, and then a restaurant would be selected based on the order history of each participant.</p><p>Once our bootcamp was over, we took a professional selfie jumped right into our first rotations.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*fzQ2qaZpruVGJ8ln" /><figcaption>Professional photo of our cohort</figcaption></figure><h3>Building for Cities</h3><p>So Bootcamp was fantastic and we learned a ton. But diving straight into PM work was like trying to play <a href="https://youtu.be/5UfPij5ABdo?t=81">Quidditch</a> after reading Harry Potter: you generally know how it works, but you’re still nowhere near ready (and maybe never will be).</p><p>I worked on the <a href="https://movement.uber.com/?lang=en-US">Cities Platform</a> team. Our work focus was to help policymakers and city planners make more informed decisions through the use of Uber’s city data.</p><p>PMs are the voice of the customer and I got to experience that first hand by organizing and leading user research interviews to understand user needs. My manager and I traveled to Seattle and London to talk to different city officials and learn about how people think about city development. These insights were important when deciding how our products should work. It hit me when I was talking about an engineer about how some data should be used and they said they wanted to lean on me because I was the one who’s actively talked to users.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*yfzYU45PO_cxX-fnn3EFWQ.png" /><figcaption>Some photos I snapped in London, Sept 2018</figcaption></figure><p>I learned that one of my core responsibilities was to gather the needs and perspectives of internal and external stakeholders and make decisions and plans based on that. <strong>Decisions should always be made based on data, not opinions</strong>. The worst thing a team could do is make a decision based on uninformed instincts or loud opinions.</p><h3><strong>Around the world in 14 days</strong></h3><p>After our first rotation, our APM class traveled across four cities to learn from local operators to understand how Uber works in different types of cities. This was undoubtedly one of the coolest parts of the APM experience.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*fNFjRo8UDROZWMyP" /><figcaption>Selfie with a few APMs near the Giza Pyramids</figcaption></figure><p>The philosophy behind the trip was that we needed to build a global solution for local problems. Everyone needs to move but how each city moves is different. As a product manager at Uber, we shouldn’t bias our product decisions based on where we’ve lived or the cultures we were exposed to. By thoughtfully understanding how each city moves, we’re better equipped to make informed decisions on a global product that serves our users.</p><p>We collectively took away so much from the trip. After the trip, we spent a month compiling our learnings and shared it out to the entire product org. The trip also gave me a lot of personal takeaways that I want to share:</p><ul><li>In Los Angeles, we saw the behind the scenes of having operational scooters across the city and learned about all of the vandalizing and thefts that happen with the different scooter parts. <strong>Building products that are both digital and physical is incredibly hard</strong> but who doesn’t love a hard problem to solve?</li><li>In Mexico City, one of the Uber drivers shared that he was in a tough spot in life and was feeling suicidal. His friend introduced him to Uber and he credits Uber for saving his life and helping him get out of his state. <strong>Hearing this person’s feelings gave meaning to our work at Uber</strong>.</li><li>In Amsterdam, we put on our courier hats on and tried to deliver Uber Eats meals across the city. We discovered a tad too late that many couriers camped at popular restaurants like McDonalds to increase their chances of getting dispatched and ended up with no dispatches. <strong>People are clever and tricks spread like wildfire.</strong></li><li>In Cairo, the road was basically a game of Mario Kart. There were bikers and cars all crammed on roads with no concept of a lane. Crossing the street as a pedestrian involved raising your hand while taking baby steps, hoping that the next car chooses not to run you over. Traffic and driving in Egypt are hectic as hell and we even got caught in a minor car accident while there. Although most cities have similar types of infrastructure components like cars, bikes, and roads, <strong>cities across the world move wildly differently</strong>.</li></ul><p>I also recorded a fun video to recap our adventure across these four cities :)</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2Fj-xkqTzNBms%3Ffeature%3Doembed&amp;display_name=YouTube&amp;url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3Dj-xkqTzNBms&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2Fj-xkqTzNBms%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/c9d8887ec932d26561632ad2735d74bb/href">https://medium.com/media/c9d8887ec932d26561632ad2735d74bb/href</a></iframe><h3>Improving the marketplace through Driver Incentives</h3><p>Once I came back from the APM trip, I started my second rotation on Driver Pricing. My team’s goal was to increase the Marketplace’s efficiency by optimizing hours that drivers are online. We accomplished this by providing <a href="http://marketplace.uber.com/pricing/driver-promotions">daily incentives</a> for the following week. By informing drivers when and where there would be incentives ahead of time, drivers can plan ahead of time and shift their hours to earn more.</p><p>Throughout the rotation, I had learnings in three areas: improving how the team collaborates, defining plans, and presenting.</p><h4>Product managers should facilitate, not relay</h4><p>Improving how the team works together took a lot of tries. It meant being very intentional with meetings and ensuring that the right people were involved. I learned that although product managers are frequently a central node for gathering information for decisions, they should not be a middle man when it comes to execution.</p><p>For example, product managers usually decide the scope of a project. But when engineers are implementing designs, product managers shouldn’t be playing a game of broken telephone between engineers and designers. Instead, they should focus on connecting the dots and make sure that the right stakeholders are talking. This feels obvious in hindsight, but hindsight is also 2020. I realized that it’s important to <strong>continuously</strong> challenge the “how” and “what” to find better ways to do things.</p><h4>Planning becomes much harder when you realize you have limited resources</h4><p>The hardest part about defining a plan wasn’t deciding what was included. It was deciding what will be excluded from the plan. There’s merit behind all ideas which is why it’s important to do brainstorms with the team during planning. When we took an initial pass at our H2 plan, it was easy to ambitiously bloat the plan. Once sizing estimates were made on all of the projects, then the real challenge began.</p><p>Prioritizing projects started with understanding our team’s strategic goals for the half. It was easier to align on a goal first than to align on projects. Once it was clear which metrics we were focused on for the half, it became easier to collaboratively cut down projects that weren’t aligned with our goals.</p><h4>Q&amp;A is more of an art than a science</h4><p>Within Marketplace, teams present their work to Marketplace leadership. This was a great way for tech leaders within Marketplace to learn about a team’s focus within a short session and ask questions. For these sessions, hard questions were a given. Given the limited amount of time, we needed a strategy to handle questions effectively.</p><p>It’s easy to get caught in a rabbit hole or go on tangents when answering questions. My skip manager and I had a great conversation about different strategies to handle difficult questions and she helped me challenge the idea of a question.</p><p><strong>Questions are usually followed by a response, but the response doesn’t have to be an answer</strong>. In other words, before answering a question, there are a lot of meta-questions in between such as:</p><ul><li>Is this question important or not?</li><li>Am I the best person to answer this question?</li><li>Does anyone in the room know the answer to this question?</li><li>If we can’t answer the question, can we follow up after?</li></ul><p>By understanding the intention and context of the question, it‘s’ easier to form the right response. Questions should aim to be resolved instead of answered. If the question is relevant and important, then the right response may be an answer. However, <strong>the right response is not always an answer</strong>.</p><h4>Understanding the engine behind Uber</h4><p>One of my favorite parts about my second rotation was my exposure to the Marketplace team. I developed a deeper interest in economics and really enjoyed thinking about the chain of effects that can happen due to product changes. Economics was no longer just a supply and demand black box concept. Instead, I got to see a lot of it in action.</p><p>Throughout my rotation, I really enjoyed learning about marginal benefits and costs and understanding how these concepts were applied to optimize driver incentives. If you’re interested in learning more about the Marketplace, you can read more about it <a href="http://marketplace.uber.com/">here</a>!</p><h3>Launching Uber Eats for Businesses</h3><p>For my last team, I worked as the first PM on Eats for Business, also known as E4B. We focused on unlocking a new business segment by providing meals for work.</p><p>When our team was first formed, we didn’t have a proper solution for employees. Uber Eats was previously designed for consumer which meant that the person ordering was always paying out of pocket. The initial problem our team faced was how could we bootstrap the existing consumer product to make it work for businesses?</p><p>The solution wasn’t trivial. There were so many different types of users, there were endless lists and sheets of feature requests from many sales reps, and there was only so much we could do in a short little time. The biggest challenge when tackling this project was managing all of the stakeholders. We needed a timeline to get an MVP out the door. This also meant establishing expectations and a timeline for when we would have things ready.</p><p>Getting a timeline was tricky. It required defining the scope of the feature and then investigating the tech effort required for the feature. Tech effort typically included design, engineering, testing, and rolling out. And this was only for one feature.</p><h4>A plan for a plan is better than a rushed estimate.</h4><p>A need for a timeline doesn’t have to be resolved with an actual timeline. Any estimates for a timeline at that point would have been purely intuitive and not based on research. Instead, we opted to provide a plan for a plan. In other words, we provided a date for when we would have a timeline ready instead of providing the actual timeline immediately.</p><p>The scenario that this avoided was if the timelines changed last minute. Stakeholders could have lost trust in our future timelines but more importantly, many customers we had relationships with may have lost trust in our team.</p><p>There’s pressure to get things out early in a competitive B2B space. Many competitors already had a competitive solution suite. I frequently heard from sales reps that if we didn’t get a feature out by a certain date, we might have lost a large deal. Our team was playing catch up with competitors and we needed to move both fast and strategically.</p><h4>The MVP is the minimal change required to solve the use case.</h4><p>But that wasn’t the approach we took unfortunately. Our MVP was defined by cutting corners from the long term vision of the feature. We were trying to make our product run at 50% before it could walk at 100%. Once we started scoping the long term engineering work, that’s when we realized that our features were not complementary. In other words, any long term work would have replaced our MVP. This would have been fine if we were a small startup, but throwaway work was not an option for us.</p><h4>Be thoughtful with how plans are made and communicated</h4><p>This was key to good stakeholder management. A good litmus test to see if you’re doing something wrong is if someone is surprised by a plan. Instead, it’s important to ensure that the right stakeholders are looped in at the right times so that they can help inform any decisions.</p><h4>Scope fast, plan better.</h4><p>We ended up spending more time scoping projects to get them out the door faster and didn’t focus enough time on getting a plan ready. Luckily, our final plan was mostly aligned with what we had scoped but we still had to cut some corners on our original scoping, leading to some lost work.</p><p><strong>The best product managers are able to sustainably maximize impact with minimal resources</strong>.</p><h4>Always talk to your users</h4><p>In March, no one knew how long quarantine would last. Should we build for this pandemic state of the world with the risk of it being over by the time we finished any features? Should we continue building for the normal world if that might be years away?</p><p>We knew so little about what was going to come in the following months because of COVID. What we did know was what our users needed as a result of COVID. After hearing this feedback, we quickly pivoted our roadmap to focus on a new product called <a href="https://www.uber.com/us/en/business/products/vouchers/?utm_source=AdWords_Brand&amp;utm_campaign=CM2038101-search-google-brand_1_-99_US-National_u4business-ss_all_acq_cpa_en_Eats_Voucher_BMM_%2Buber+%2Beats+%2Bvouchers_kwd-407583000398_471259155349_109354106574_b_c&amp;campaign_id=11263602467&amp;adg_id=109354106574&amp;fi_id=&amp;match=b&amp;net=g&amp;dev=c&amp;dev_m=&amp;ad_id=471259155349&amp;cre=471259155349&amp;kwid=kwd-407583000398&amp;kw=%2Buber+%2Beats+%2Bvouchers&amp;placement=&amp;tar=&amp;gclid=Cj0KCQiAqo3-BRDoARIsAE5vnaIEZyZSOEYK8RQ5U3Jut1hVdwnq0hjgBqldacwiQGSfwqOL-6GFdAIaAuBoEALw_wcB&amp;gclid=Cj0KCQiAqo3-BRDoARIsAE5vnaIEZyZSOEYK8RQ5U3Jut1hVdwnq0hjgBqldacwiQGSfwqOL-6GFdAIaAuBoEALw_wcB&amp;gclsrc=aw.ds">Uber Vouchers for Eats</a>.</p><p>This product saw immediate product market fit and caused business growth immediately after launch. This was because we listened to our customers’ needs and our solution solved their problems.</p><h4>Find the perfect balance of urgency and calmness</h4><p>Bugs happen all the time. There will be dozens of Uber sales people trying to escalate their issues or feature requests and try to get us to prioritize their requests. In their defense, it’s their job to establish a good relationship with their clients and get any problems fixed as soon as possible.</p><p>It‘s hard to ignore my biases. My first reaction was to always try to resolve the situation immediately, listening to the loud voice or the request from a familiar person. I’ve found that challenging the request through a series of questions works best for me:</p><ul><li><strong>Is this actionable in the next year?</strong> As a product team, we always have an opinion on how the product should work and we need to make sure that the request isn’t just noise.</li><li><strong>Does this need to happen now? </strong>There’s a high cost to pulling in engineers from their existing work. It’s always preferred to delay, deprioritize or delegate the request if possible.</li><li><strong>What is the exact ask?</strong> The more information, the better. Prioritizing a vague request will only lead to building something unused.</li></ul><h4>Nothing is free.</h4><p>To work on something was to say that that was the most useful way to spend my time right now. So what if someone asks if we can tackle a new project? Saying yes to a new project meant stopping a current project. Saying no to a new project meant missing out on a new opportunity.</p><p>In other words, there’s opportunity cost behind every decision. I initially found that saying no to new requests was hard. I learned from peers that there are better ways to reframing new requests. For example:</p><p>Should we stop working on this existing thing and start working on this new thing? In other words, <strong>is this new thing more important than what we’re currently working on?</strong></p><p>If the new request is less important, then there’s that should resolve the request. We live in a world where time is limited so we need to be intentional with how our team’s time is spent. Or in economic terms, the marginal benefit isn’t worth the marginal cost!</p><h4>Having fun with product management!</h4><p>Product management is as much an art as a science. Since it’s also an art, the many degrees of freedom enables ways to have fun on the job!</p><p>For me, that meant organizing team socials and celebrations, driving blue-sky brainstorming sessions, and designing conceptual mockups. For some weekly meetings that are at unpleasant hours like Monday morning, I liked to ask fun questions about people’s personal lives such as which shows they’ve been watching on Netflix or any movies that they watched. This was a fun way to get people engaged and warmed up for the meeting.</p><h3>The end of my Uber ride</h3><p>My time at Uber has been nothing short of incredible. The team at Uber was truly special and I couldn&#39;t have imagined a better team of folks to work with for my first job. It’s been incredible working with my managers and teammates to shape Uber’s products for millions of people.</p><p>As my Uber chapter closes, I’m embarking on a new adventure. I’ll be going back to being an engineer and joining a startup in the creator space. If you’re interested in joining along for my ride, come tag along <a href="http://andyzhang.substack.com/">here</a>.</p><p>Thanks for reading!</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*LM8Uximp3MBgRE0zS5J_RA.png" /><figcaption>Sneak peek of my next adventure 👀</figcaption></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=ce9fccf1c66f" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[My Journey into Product Management]]></title>
            <link>https://andyzhang.medium.com/my-journey-into-product-management-69c3ae9200de?source=rss-ebfbfc45654f------2</link>
            <guid isPermaLink="false">https://medium.com/p/69c3ae9200de</guid>
            <category><![CDATA[tech]]></category>
            <category><![CDATA[product-management]]></category>
            <category><![CDATA[product]]></category>
            <category><![CDATA[software-development]]></category>
            <category><![CDATA[interview]]></category>
            <dc:creator><![CDATA[Andy Zhang]]></dc:creator>
            <pubDate>Wed, 30 Sep 2020 18:54:31 GMT</pubDate>
            <atom:updated>2020-09-30T18:54:31.477Z</atom:updated>
            <content:encoded><![CDATA[<h4>How a mecha-robot anime led me to becoming a product manager</h4><p>My journey into tech and product management started with an anime about robots called <a href="https://www.youtube.com/watch?v=Xb69eSq5GE4&amp;ab_channel=NGKY">Gundam Seed</a>. After every high school summer, I indulged myself with a rewatch of the series. I was obsessed with “Coordinators”, an advanced race with the ability to pilot mecha-robots. My obsession eventually led me to join my school’s robotics team as a programmer. I wanted to be a robot programmer.</p><p>After a couple of physics courses that explained the impossibilities of mecha-robots and some humbling bugs, I shifted my focus into software engineering.</p><h4>University of Waterloo opened my eyes to the world of tech</h4><p>That fall, I started as a freshman at the University of Waterloo’s Software Engineering program. My college was known for its unique co-op program that required students to complete internships as a part of their degree. In my case, we were expected to complete six internships. This meant that students would graduate with two years of work experience across up to six different companies.</p><p>I started my stint in the tech industry as a software engineering intern at Facebook. I was lucky enough to land an internship after getting a foot in the door as the Waterloo Facebook Hackathon winner in 2014.</p><p>Following Facebook, I continued to work as a software engineer by interning at Google and then Coursera. Throughout these internships, I was exposed to various types of problems. For example, at Facebook, I focused on increasing user engagement with Facebook Pages. At Google, I built an internal tool to analyze Chrome usage dat to understand how Google Chrome was being used which helped inform a redesign. At Coursera, I implemented Android designs for new product called specializations and improved the translation processes.</p><p>At the end of the day, <strong>something felt missing</strong>. I didn’t feel as fulfilled with the work that I was doing. At the end of the day, I got fulfillment by solving user problems.</p><p>In freshman year, I always craved pizza but didn’t want to spend money. So I wrote a script to pull all of the <a href="http://www.ceca.uwaterloo.ca/students/sessions.php?month_num=9&amp;year_num=2013">campus’ employer info sessions</a> into my calendar so I knew where to get free pizza. To me, eating that free pizza was fulfilling because it meant I had solved the user’s problem i.e. I craved pizza.</p><h4>How do I get more “free pizza”?</h4><p>Problem solving through mini side projects was my first step into product. At the time, I didn’t know what it meant to “get into product”. The path towards product felt nebulous. At the time, I thought the closest role to a user was a product designer because it had “product” in the title. My brute force approach to getting into product led me into the world of product design.</p><p>I spent most of my leisure time trying to understand product design and exposing myself to it. In my junior year, I designed for Hack the North and interned as an engineer at Figma. Figma was just over 10 people at the time and filled with incredibly smart and interesting people. Closely working with the Figma team was a transformative experience. It taught me how to be a better teammate, how to work with designers, how products are designed and how important culture is.</p><p>After Figma, I wrapped up my school’s co-op requirements with a product design internship at Quora and a PM internship at Microsoft. Getting my feet wet with product throughout these internships helped me understand gave me a sense of how designers and PMs fit into a cross-functional tech team. However, since product lifecycles frequently outlast an internship, I merely scratched the tip of the iceberg.</p><p>As I ended my six internships, a lingering interest in product pushed me into interview prep for product manager. <strong>I was excited about the prospect of solving difficult problems that could improve people’s lives</strong>, whether it was through giving people free pizza or helping them get from point A to point B.</p><h4>Landing a product manager job</h4><p>Working full-time is daunting. The idea of trying to get a full-time job is daunting. I wanted to be a product manager but it wasn’t clear how I could make the leap into a full-time product management role. <strong>I still struggled to answer the question “what does a product manager do?”</strong></p><p>My background in software engineering led me to believe that I just needed to find the LeetCode equivalent for PMs. Spoiler alert, there’s no LeetCode equivalent. However, there is a Cracking the Coding book equivalent.</p><p>Cracking the PM interview, a book also authored by Gayle Laakmann, is a great book to get started with PM interviews. Its contents reflect the books title: it teaches you what to expect from PM interviews and how to prepare for them.</p><p>However, I’ve found that PM interviews feel a lot less predictable than coding interviews. In coding interviews, problems revolve around systems that have deterministic outputs. In PM interviews, there are a lot more unpredictable factors such as people, businesses, competitors, and much more.</p><p>In addition, the solution an interviewer looks for is also much broader than a coding interview. In other words, there’s no such thing as a perfect answer in PM interviews.</p><h4>Exposure is the best way to becoming better at PM interviews</h4><p>PM interviews are a bit more of an art than a science. Unfortunately, it’s not possible to get away with a hash map equivalent for PMs and call it a wrap.</p><p>To improve, one just needs exposure, practice, and feedback. Practice can be done with friends and feedback can be done with self-retrospectives. Exposure is a piece that’s always been difficult to get. The first time I was exposed to how other people approach PM interviews was when I was assisting another interviewer.</p><p>After conducting several interviews, I realized that a lot can be learned from observing other people approach interviews. Similar to how you can learn years of experience through a book, it’s possible to learn new ways of thinking from years of problem solving in a matter of minutes.</p><p>I recently worked with <a href="https://www.productalliance.com/">ProductAlliance</a> to help provide content that can hopefully help people interested in product management become PMs. <strong>One thing that makes their course stands out is that they provide an inside scoop into how many veteran PMs approach classic interview problems that most aspiring PMs will face</strong>. If you’re an aspiring PM, ProductAlliance will not only provide a structured approach to cracking the PM interview, but it will also upgrade your problem solving toolkit to nail your next PM gig.</p><h4>Any non-PM experience is valuable as a PM</h4><p>What makes PM super exciting is that it’s a discipline that involves bringing a unique problem solving perspective. People don’t learn how to become PMs, they learn how to apply their perspectives into PM. That’s the reason why most PMs come from so many different backgrounds.</p><p>My past experience as an engineer and designer has offered me a unique perspective on each problem I solve. That makes me a better PM for certain types of teams, but not necessarily a bad PM for other teams.</p><p>If you’re an aspiring PM and you have a coding background, great! PM is just around the corner! If you’re an aspiring PM but don’t have a coding background, the road is far from being closed off! You’ll need to get slightly creative and take a detour, but it’s not much harder than starting off with a coding background.</p><p>My personal journey into product management was challenging. The detours along the way were bumpy but they’ve made me a better product manager because of the perspectives they’ve given me. If you’re interested in learning more about product management, come chat with me <a href="https://jemi.app/andy">here</a>. Thank you for taking time out of your day to read my journey!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=69c3ae9200de" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Data-Driven Introspection]]></title>
            <link>https://andyzhang.medium.com/data-driven-introspection-e2d30ef41181?source=rss-ebfbfc45654f------2</link>
            <guid isPermaLink="false">https://medium.com/p/e2d30ef41181</guid>
            <category><![CDATA[introspection]]></category>
            <category><![CDATA[data-analysis]]></category>
            <category><![CDATA[data]]></category>
            <category><![CDATA[travel]]></category>
            <category><![CDATA[self-improvement]]></category>
            <dc:creator><![CDATA[Andy Zhang]]></dc:creator>
            <pubDate>Thu, 27 Sep 2018 18:52:29 GMT</pubDate>
            <atom:updated>2018-09-27T18:57:16.598Z</atom:updated>
            <content:encoded><![CDATA[<p>Earlier this year, I embarked on a month-long solo grad trip to East Asia. The idea of being socially disconnected fascinated me, and the freedom of owning each hour of my day was liberating. It was the ideal environment for introspection, and for reflection.</p><p>I’ve equally been fascinated by the idea of leveraging data to discover something new. Data is a step forward towards answering the unanswered questions.</p><p>And so the process starts with this question: <strong>What did I want to learn about myself?</strong></p><p>If you’re interested in learning about the methodology, you can find it at the end of this post. Otherwise, here are 4 of my main takeaways:</p><h4>1. I am 22% happier on average during days where I hang out with friends</h4><p>Throughout my trip, I met up with friends who were around the same area. This insight is based on an arbitrary scale from 1 to 5. Despite being arbitrary, a t-test shows that the difference was statistically significant(p-value=0.005), meaning that I’m significantly happier when I see friends.</p><p>It’s worth noting that this result is an average. I’ve had bad days with friends, and good days alone. This insight does suggest that I’m an extrovert on an average day.</p><h4>2. My iPhone camera usage time is correlated with my happiness</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*1mnjUvNGHYXrbuK1NWY32A.png" /><figcaption>Shot on iPhone X at Miyajima Island</figcaption></figure><p>Although a causation is unlikely, there are several plausible explanations for this. If I’m having a good time, I may have had a tendency to capture a photo of the moment. After all, it’s nice to relive those moments through a photo sometime in the future. My future self will thank me for that.</p><p>Alternatively, using my iPhone camera is fun, and fun things make me happy. This wouldn’t be surprising given that I enjoy taking photos of <a href="https://www.instagram.com/andy.fuji/">people and places</a>.</p><h4>3. Coffee and other diuretics cause me to use the restroom 21% more</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*nz61jSKenuH0-pAE.jpg" /></figure><p>Everybody’s bodies react differently to diuretics. I didn’t understand the detailed effects on my body, so I was curious to see through data how it could affect my day. Based on the results, it turns out that it does have a significant affect on my habits during the day.</p><h4>4. Less Instagram = More Reddit</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/399/1*9PDERCepmr4E0qaD73euQw.png" /><figcaption>Reddit and Instagram usage pattern</figcaption></figure><p>One of my goals for my grad trip was to reduce my social media consumption. This meant cutting down on Facebook, Instagram, and more. It’s a known fact that habits can only be replaced. Here’s that fact shown above in data.</p><p>To clarify, Reddit wasn’t the only thing that I spent my time doing. I often took breaks at coffee shops to read books, Quora, or sipping on a coffee. That said, prior to my grad trip, I rarely used Reddit which was why the increased usage was so interesting.</p><h4>And so much more…</h4><p>There’s a plethora of other types of analysis that can be done on the data, which made this project all the more fun. Although these insights don’t provide a comprehensive answer to the original question, it’s a step forward.</p><h3>Methodology</h3><p>To answer my original question, I needed to collect data. A week prior to my trip, I wrote down a list of measurable things about myself. For example:</p><ul><li>How happy I felt (from 1–5)</li><li>How many minutes I spent on X app</li><li>How many times I used the washroom</li><li>How much did I spend</li><li>Etc.</li></ul><p>Each day during my trip, I summarized the day’s data into a spreadsheet. Here’s an example of what my app usage spreadsheet page looked like:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*KqwtPg7eMVZsHaJ4bd3qeA.png" /><figcaption>App usage over time in minutes.</figcaption></figure><p>At the end of my trip, I processed all of this data to answer the aforementioned question using Jupyter Notebooks. To calculate significance, I ran one-way ANOVA’s or a Pearson correlation coefficient over 38 samples (38 days).</p><p><em>Thank you for reading! If you enjoyed reading this piece, I’d love to hear your thoughts in the comment section below! If you’re interested in seeing my grad trip as a short video, check it out </em><a href="https://youtu.be/WU7FJy_zh7Q"><em>here</em></a><em>!</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e2d30ef41181" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[My Favourite Project during University]]></title>
            <link>https://andyzhang.medium.com/my-favourite-project-during-university-1e0ba5ad4bf9?source=rss-ebfbfc45654f------2</link>
            <guid isPermaLink="false">https://medium.com/p/1e0ba5ad4bf9</guid>
            <category><![CDATA[university]]></category>
            <category><![CDATA[side-project]]></category>
            <category><![CDATA[data-science]]></category>
            <dc:creator><![CDATA[Andy Zhang]]></dc:creator>
            <pubDate>Wed, 18 Jul 2018 00:00:00 GMT</pubDate>
            <atom:updated>2018-07-23T18:15:04.194Z</atom:updated>
            <content:encoded><![CDATA[<p>After seeing the Syste<a href="http://joeyloi.com/SYDE2017classprofile.pdf">ms Design Engineering 2017 Class Profile</a>, I was deeply inspired to create a class profile for my Software Engineering class. I owe a lot to my class and my program, and felt incredibly lucky to be a part of it. Prior to university, I knew very little about software engineering, so studying it at the University of Waterloo never occurred as an option. If my <a href="https://www.youtube.com/channel/UCV0qA-eDDICsRR9rPcnG7tw">friend</a> never suggested to me to study at the University of Waterloo, I wouldn’t be where I am today.</p><p>I wanted to create a class profile capturing what I wished I knew before university. I wish I knew what studying software engineering was like, what opportunities were available, how much debt I’d graduate with, if it’s too late for me to start coding, etc. There’s a lot of focus from the media on what software engineers in the industry are like, but the path connecting their beginnings to employment is still mysterious.</p><p>The goal of our class profile is to share the life of software engineering students, covering topics such as co-op, academics, lifestyle, and much more, and to increase the transparency of the transition from a high school student to a full-time software engineer.</p><p>You can check out the final result <a href="https://classprofile.andyzhang.net/">here</a>.</p><p><a href="https://classprofile.andyzhang.net/">UW Software Engineering 2018 Class Profile</a></p><h3>Understanding the questions</h3><p>I gathered what people wanted to know about Software Engineering (SE) students through a survey. The goal of the survey is to gather questions from the target audience. In this project, the target audience is composed of students who are either interested in or new to software engineering.</p><p>I created a Google Form to collect questions current SE students wish they knew during high school, or what they wanted to learn more about. SE students were asked because they were students who were once interested in SE and are possibly still new to SE (depending on the year). In total, I collected 152 questions from 40 survey responses across 5 class of SE students.</p><p>I went through all of the questions and collapsed any questions that were redundant. I then classified the question type:</p><ul><li>One dimensional: Can be answered by one data point. Ex: How many hackathons did students attend?</li><li>Two dimensional: Can be answered by two data points. Ex: Do grades affect salary?</li></ul><p>Afterwards, I removed questions that were too specific or could not be included in the survey (dating life, sex, etc.). The latter questions were separated into a second anonymous class survey and is not associated with this class profile.</p><h3>Designing the survey</h3><p>One of the main challenges was trying to get a high response rate. Since the survey involved sharing personal data, students were more likely to be reluctant to respond. The goal was to get an 80% response rate, so I made several compromises to the survey:</p><ul><li>It should take 15 minutes to respond, meaning we had to cut questions.</li><li>Its responses should not be seen by any students including myself. The technicalities are discussed in the section below.</li></ul><p>Once the questions were fully sorted, the questions were broken down into individual survey questions. Some questions were also shortened for brevity. For example, hackathon attendance per term was shortened to hackathon attendance throughout university.</p><p>I did a beta test on a small group of students to gather feedback on the survey. The feedback consisted of catching edge cases (ex: there’s no numerical grade during an exchange term, some people skipped a co-op). After addressing the feedback, the class survey was ready.</p><h3>Surveying the SE class of 2018</h3><p>To further incentivize the class to fill out the survey, I talked to several professors to see how they’ve incentivized students in the past to fill out faculty surveys. Luckily, Derek Rayside, our FYDP professor, was supportive of the class profile and offered to add marks for everyone if the survey reached an acceptable response rate.</p><p>In addition, Patrick Lam, the director of Software Engineering, agreed to teach our weekly class wearing a silly outfit if we could reach an 80% response rate.</p><p>The survey closed with a 79.6% response rate.</p><h3>Working with anonymous data</h3><p>Processing the data became fairly tricky because the survey data was on Patrick Lam’s computer. This was an important constraint because many students expressed concerns on their personal data and didn’t want it to be seen by any students including myself. As a result, Christopher Luc and I came up with a system to process the data anonymously.</p><p>Before programmatically processing the data, the data needed to be normalized. For example, Snap, Snapchat, Snap Inc becomes Snap.</p><p>To view a column without being able to infer which row it belonged to, the system shuffled each column independently from each other. Once each column was shuffled, it created a local file storing the shuffled column as well as the original row indices. These indices were a key piece to rebuilding the original data.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/761/0*RCzY-nNTbmp4l80f.png" /></figure><p>Patrick only sent me the shuffled columns and held on to the original indices. I manually normalized the shuffled data, fixed any typos and ensured that answers were properly formatted so that they could be processed programmatically.</p><p>Once I finished normalizing the whole dataset, I sent back the shuffled columns so that Patrick could stitch each normalized shuffled columns into a new dataset.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/761/0*bA-2X_8sjYMR6lSP.png" /></figure><h3>Transforming the data</h3><p>There were two types of findings that we were looking for: one dimensional (how many X in the class?), and two dimensional (does X affect Y)? I already had the data I needed for the one dimensional findings, but needed to group data for the two dimensional findings.</p><p>I wrote several scripts to transform the raw data into a structured format that didn’t reveal anything about a student. This was done by minimizing the coupling between each column of a row. For example, if I wanted to see the effect of grades on salary, I grouped each grade with the corresponding salary into a pair. Each of these pairs were disconnected from the pairs from other terms.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*biowQlhSvl1Izn3B.png" /></figure><p>In this example, disconnecting data from each term prevented me from identifying a recognizable salary (i.e. outliers), and then reading the row. Since the terms were disconnected, I could not see the data from the person’s other terms.</p><p>These scripts were shared with Patrick Lam so that he could run them on the normalized data sets. He then sent me back the structured datasets for me to process.</p><h3>Learning from the data</h3><p>After I finished transforming the data, I visualized each data set to determine whether there was something interesting in the data. Since the structure of each data set was similar, I was able to reuse my visualizations easily, enabling me to figure out what had potential.</p><p>Although visualizations didn’t help me figure out if something was statistically significant, it did help me filter out what was not statistically significant.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*aN8nNa_888zNbR35.png" /></figure><p>In this example, it was easy to tell that there was a low chance of there being a correlation between parental education and salary per term.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*WoZeKsQdyYtOooKA.png" /></figure><p>In this example, I was plotting hackathon attendance vs salary. I noticed that there was a difference between students who didn’t attend hackathons and students who did. This led me to running more rigorous statistical tests to see if there was a correlation.</p><h3>Statistical verification</h3><p>After visualizing all of my transformed data sets, I narrowed my focus down to the data sets with potential. I consulted with Christopher Luc, a two time Data Science intern (meaning he’s legit!), what methods he would recommend for the data I had. There were two methods I used: one way ANOVA, and evaluating correlation coefficients.</p><p>I ran the one way ANOVA on data sets involving groups of distributions. For example, gender vs salary had two groups. Each entry from a student was their average salary. The aggregation of each average salary led to a distribution for both male and female students.</p><p>I evaluated the one way ANOVA p-value from the groups to determine if there was a significant difference. If the p-value was lower than 0.05, it means that there is a significant difference between the groups. In other words, it meant that there was a dependent variable who’s value is affected by the independent variable’s value. In most cases, grades and salary were the dependent variables.</p><h3>Designing the page</h3><p>As I wrote the report, I borrowed a lot of inspiration from other reports such as Systems Design Engineering (SYDE) 2017 Class Profile, HackerRank’s Developer Skills Report, and more. The report had to be representative of what brought our class shared in common, and that was software engineering.</p><p>I wanted the report to be readable, but also resemble to a command line, which is a software engineer’s best friend. I settled with a monospace and a sans serif font pairing for the content. I borrowed colours from my favourite terminal colour schemes such as Solarized, and also handpicked several desaturated colours that complimented the colour scheme.</p><p>For the cover photo, I wanted to have things that were common in the daily lives of software engineers. This included coffee, code, web browser, laptop, servers, etc. I put them all together into an isometric-styled illustration, and applied a suitable blue and purple colour gradient on it. Purple is a colour that represents engineering in Canada, and blue is a colour common in tech companies (Facebook, Dropbox, Twitter, Dell, IBM, etc.)</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*jHq8Cct_sqHdD8nV.png" /></figure><h3>Building the website</h3><p>It seemed fitting to build the class profile given it’s about a software engineering class. One of the main priorities for the class profile was that it was a good reading experience across platforms. Therefore, it needed to load quickly, and also be responsive for mobile.</p><p>To reduce load time and increase performance, the use of images was minimized. Although creating images of charts would be straightforward with existing native software, I built the visualizations using D3 in order to reduce load time as well as optimize for responsiveness. Building the visualizations from scratch also allowed reusability and better customization.</p><p>The final website had a total bundle size of 1.5MB. In contrast, Reddit’s homepage has a total bundle size of ~1.7MB.</p><h3>Launch the class profile</h3><p>The launch date was set to be at the beginning of a week on a Monday under the assumption that more people will read this on a weekday than on a weekend. That way, the lifespan of the content will last a full 5 days rather than get cut by the weekend. I was hoping that it would also be a great conversation topic for the week for readers, allowing it to gain more potential traction.</p><p>I introduced the class profile through a Medium post, similar to the SYDE 2017 Class Profile. This showcased some of the highlights from the class profile, and linked to the full class profile should the reader want to learn more.</p><p>Since prospective students were the target audience, I also worked with the Software Engineering department to include a link to the class profile on the program’s website. This helps prospective students learn more about what studying software engineering is like.</p><h3>Reflecting on my favourite project</h3><p>This was one of the most challenging but exciting projects I’ve worked on. Nothing excites me more than seeing so many different pieces come together into one final result. There’s still so much I would want to include in this post, but I didn’t want to make this as long as the class profile itself. The process involved a wide range of responsibilities such as research, talking to stakeholders, engineering, design, data science, etc., all of which are responsibilities that I deeply enjoyed.</p><p>There was something new to learn in each part of the process. For example, learning how to make isometric illustrations, familiarizing myself with Jupyter notebook, studying statistical tests, and more. Applying these learnings was just as fulfilling as learning them, making the process enjoyable from start to finish.</p><p>Thank you Christopher Luc for reviewing the draft of this post. Chris was integral in designing many parts of the process and pointing me in the right direction for statistical tests. Learn more about what he’s up to on his <a href="https://medium.com/@chrisjluc">Medium</a>!</p><p>If you’re interested in learning more about our class profile, I’d love to chat! You can reach out to me via <a href="/service/mailto:andy.zhang@edu.uwaterloo.ca?subject=Hello!">email</a> or say hi on <a href="https://twitter.com/andyzg3">Twitter</a>!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=1e0ba5ad4bf9" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[University of Waterloo Software Engineering 2018 Class Profile]]></title>
            <link>https://medium.com/uwaterloo-voice/university-of-waterloo-software-engineering-2018-class-profile-aae48c2075a0?source=rss-ebfbfc45654f------2</link>
            <guid isPermaLink="false">https://medium.com/p/aae48c2075a0</guid>
            <category><![CDATA[data]]></category>
            <category><![CDATA[university-of-waterloo]]></category>
            <category><![CDATA[data-visualization]]></category>
            <category><![CDATA[surveys]]></category>
            <category><![CDATA[software-engineering]]></category>
            <dc:creator><![CDATA[Andy Zhang]]></dc:creator>
            <pubDate>Mon, 18 Jun 2018 15:13:43 GMT</pubDate>
            <atom:updated>2019-01-17T00:08:03.185Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*4nEHd2jjXQbuEMDcdiM7WQ.png" /></figure><p>After studying 5 years of software engineering, I’ve learned that there’s no such thing as “too much documentation”. As many software engineers would agree, documentation is considered a best practice.</p><p>Inspired by the Systems Design Engineering 2017 Class Profile, the Software Engineering Class of 2018 worked together to compile our data to document who we were and who we’ve become. We also wanted a deeper understanding of the factors that brought students to where they are today, whether it’s how many side projects they’ve built or their family backgrounds.</p><p>In collaboration with the Software Engineering faculty, we’ve put together this data into a comprehensive class profile with 113 (out of 141) students. Check out the highlights below, or <a href="https://classprofile.andyzhang.net/">click here</a> to view the full version.</p><h3>Co-op opened many doors for students</h3><p>As seen in the visualization, a large fraction of the class had the opportunity to work for well known tech companies such as Facebook, Microsoft, Google, Amazon and so much more!</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*_ffto0CTMiMKUethgU201g.png" /><figcaption>Companies students have worked at</figcaption></figure><p>Some students also used their co-op terms to explore different career paths, such as product design and data science. John Salaveria, an avid illustrator and fashion blogger, leveraged his software engineering experience with his design interests to join a “community of [product] designers with passions similar to [his] own.”</p><h3>65% of students are graduating debt-free.</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*rw4YEACdiH4GmKIA8JvhQg.png" /></figure><p>Thanks to co-op, most students were able to finance their university education. It’s important to note that there were students from each family income bracket who graduated debt-free. Among those who graduated in debt, 85% of them had less than $30,000 of debt.</p><h3>By the last co-op term, salaries tripled in comparison to the first.</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*dgr8NbyrrdH3R1Qb1hsNCw.png" /><figcaption>Compensation for each co-op term</figcaption></figure><p>Students can only hope that this trend continues and that their salary triples again in 5 years. Who knows? (…it probably won’t happen)</p><p>Obviously, students have more experience in their later terms, making them eligible for higher paying jobs. However, other factors help explain this increase, mainly higher living costs. Employers compensate for higher living costs with housing stipends. For example, San Francisco housing typically costs <strong>$1,500-$2,500 USD a month</strong>, which is enough to cover a <strong>full 4 months</strong> in Waterloo.</p><h3>Good grades do not imply higher job pay</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*DMd0QiS1kfwoUVKWbGwe_w.png" /><figcaption>Green: 1st co-op, Orange: 6th co-op. Correlation coefficient: 0.23, -0.05.</figcaption></figure><p>There was no correlation found between grades and co-op salary. This finding is applicable to all co-op jobs. Instead, there are more important factors that affect job pay such as side projects, extracurricular activities, and previous relevant experience.</p><h3>Learn more!</h3><p>If you enjoyed these highlights, there’s a lot more content like this! You can view the full class profile at: <a href="https://classprofile.andyzhang.net/">classprofile.andyzhang.net</a> and read content covering co-op, student background, school, program outcome, and what’s next for students.</p><p>I hope you enjoy reading our class profile! Let me know what you think in the comments, or drop me a line at <a href="http://mailto%40andy.zhang@edu.uwaterloo.ca/?subject=Class%20Profile">andy.zhang@edu.uwaterloo.ca</a>.</p><p><a href="https://classprofile.andyzhang.net/">UW Software Engineering 2018 Class Profile</a></p><p><em>Important note: (July 28, 2018) Some numbers were corrected after discovering an inconsistency in the data; impacts were at most 5% except for the max salary numbers, which changed from 20k to 16k.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=aae48c2075a0" width="1" height="1" alt=""><hr><p><a href="https://medium.com/uwaterloo-voice/university-of-waterloo-software-engineering-2018-class-profile-aae48c2075a0">University of Waterloo Software Engineering 2018 Class Profile</a> was originally published in <a href="https://medium.com/uwaterloo-voice">uWaterloo Voice</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Keyboard Shortcuts ⌨ ❤️]]></title>
            <link>https://andyzhang.medium.com/keyboard-shortcuts-%EF%B8%8F-697cece49d4?source=rss-ebfbfc45654f------2</link>
            <guid isPermaLink="false">https://medium.com/p/697cece49d4</guid>
            <category><![CDATA[productivity]]></category>
            <category><![CDATA[keyboard]]></category>
            <category><![CDATA[shortcuts]]></category>
            <dc:creator><![CDATA[Andy Zhang]]></dc:creator>
            <pubDate>Fri, 11 Nov 2016 02:33:23 GMT</pubDate>
            <atom:updated>2016-11-11T18:46:34.203Z</atom:updated>
            <content:encoded><![CDATA[<p>Shortcuts are amazing. I’m trying to get better with useful shortcuts in all of my favorite products as well as my OS(macOS).</p><p>Here’s a list of my favorite shortcuts on some of the applications I use most often. There are only so many I know, so if you have any you’d like to share, please add them as a comment, <a href="https://twitter.com/andyzg3">tweet me</a>, or <a href="/service/mailto:andzhng@gmail.com">email me</a> your favorite keyboard shortcut.</p><h3>macOS</h3><h4><strong>Text Deletion</strong></h4><ul><li>Fn+Delete: Erase one character right of the text cursor</li><li>Option+Delete: Erase from text cursor to previous word ← my favorite</li><li>Cmd+Delete: Erase from text cursor to beginning of line</li></ul><h4><strong>Text Navigation</strong></h4><ul><li>Cmd+(←, ↑, →, ↓): Move text cursor to the extreme of that direction</li><li>Option+(←, →): Move text cursor back/forward one word</li><li>Option+(↑, ↓): Move up to beginning of previous line/end of next line</li></ul><h4><strong>Other</strong></h4><ul><li>Cmd+Z: Undo</li><li>Cmd+Shift+Z: Undo the undo ← a life saver</li><li>Cmd+Z → Cmd+Shift+Z: …Nothing :)</li><li>Ctrl+Cmd+Space: Emojis 💩</li></ul><h4><strong>App Navigation &amp; Management</strong></h4><ul><li>Cmd+Tab: Switch apps</li><li>Alt+Tab: Switch windows. Hold shift to go in reverse direction</li><li>Cmd+~: Switch windows of the same app ← super useful</li><li>Cmd+M: Minimize window(only way to get it back is from the Dock)</li><li>Cmd+H: Hide window</li><li>Cmd+W: Close window</li><li>Cmd+Q: Close all windows of this app and end the apps process</li><li>Cmd+, : Settings page</li></ul><h3>Apps &amp; Websites</h3><h4>Chrome Tab Navigation &amp; Management</h4><ul><li>Ctrl+Tab: Next tab</li><li>Ctrl+Shift+Tab: Previous tab</li><li>Cmd+(1–8): Go to 1st-8th tab ← huge time saver</li><li>Cmd+9: Go to last tab</li><li>Cmd+L: Go to URL bar</li><li>Cmd+(←, →): Go to previous/next page</li><li>Cmd+(↑, ↓): Go to top/bottom of page</li><li>Cmd+W: Delete tab</li><li>Shift+Click a different tab: Select multiple tabs. Cmd+W: Close multiple tabs</li><li>Cmd+Shift+T: Open last closed tab</li></ul><h4>iTerm2</h4><ul><li>Ctrl+R: Reverse search a command ← one of the coolest</li><li>Ctrl+A or Fn+←: Beginning of line</li><li>Ctrl+E or Fn+→: End of line</li><li>Ctrl+U: Erase the whole command</li></ul><h4>Slack</h4><ul><li>Cmd+(1–9): Switch to the (n)th team</li><li>Cmd+K: Jump to a conversation(channel or direct message)</li><li>Option+(↑,↓): Go up/down a channel from the selected channel</li><li>Shift+Option+(↑,↓): Go to the previous/next unread channel</li><li>Option+Click: Mark message as unread</li></ul><h4>Messenger</h4><ul><li>Option+(↑,↓): Go to the previous/next conversation ← mindblown</li></ul><p>—</p><p>Shortcuts recommended by readers:</p><h4>Screenshots</h4><ul><li>Cmd+Shift+4: Drag screenshot</li><li>Cmd+Shift+4 then Space: Screenshot of a window</li><li>Cmd+Ctrl+Shift+4: Screenshot into clipboard</li></ul><h4>Finder</h4><ul><li>Cmd-Shift-C: Go to Computer Directory</li><li>Cmd-Shift-H: Go to Home Directory</li><li>Cmd-Shift-A: Go to Applications Directory</li><li>Cmd-Shift-Delete: Empty Trash</li></ul><h4>Chrome</h4><ul><li>Cmd+Option+J: Opens console</li><li>Cmd+Option+I: Inspect Element</li><li>Cmd+Q note: Click Chrome on top right, and “Warn before quitting”</li><li>Cmd+Shift+N: New Incognito Window.</li><li>Vimium(Chrome Extension): Vim bindings for Chrome.</li></ul><h4>macOS</h4><ul><li>Opt+Shift+(Volume/Brightness keys): Changes volume/brightness in smaller increments</li><li>Opt+(Volume/Brightness keys): Opens System preferences for the respective setting.</li><li>Cmd+Shift+([, ]): Prev/Next tab on apps(works in Chrome, iTerm…)</li></ul><h4>iTerm</h4><ul><li>Ctrl+W: Delete word before cursor</li><li>Tab: Auto complete file and folder name</li><li>Ctrl+Z: Puts whatever you’re doing in a suspended background process. fg restores it.</li></ul><p><em>Thank you to Shane Creighton Young, Arash Rai, Xenia Tay, Kevin Bedi, Justin Sacbibit, Jonathan Ma, Shahmeer Navid, Toby Thomas and Paul Bardea for your contributions.</em></p><p>There are probably more useful shortcuts that I don’t know and would love to learn from you. Hope you learned a new shortcut!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=697cece49d4" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Building a kickass idea at Battle of the Hacks]]></title>
            <link>https://andyzhang.medium.com/hack-the-north-at-battle-of-the-hacks-eb5b2333c9c2?source=rss-ebfbfc45654f------2</link>
            <guid isPermaLink="false">https://medium.com/p/eb5b2333c9c2</guid>
            <category><![CDATA[hackathons]]></category>
            <category><![CDATA[mobile-app-development]]></category>
            <category><![CDATA[mobile]]></category>
            <dc:creator><![CDATA[Andy Zhang]]></dc:creator>
            <pubDate>Wed, 20 Jul 2016 04:55:47 GMT</pubDate>
            <atom:updated>2016-07-20T05:08:19.418Z</atom:updated>
            <content:encoded><![CDATA[<p><strong><em>Hackathon:</em></strong><em> A 24–48 hour event where teams of up to four build something awesome. Examples include </em><a href="http://www.cbc.ca/news/canada/kitchener-waterloo/pokemon-go-similar-to-game-students-created-at-uw-competition-in-2014-1.3676833"><em>a 2014 game similar to Pokemon Go</em></a><em>, </em><a href="http://devpost.com/software/silicon-man"><em>Iron Man’s hipster tech cousin</em></a><em> and more.</em></p><p>—</p><p>A few weeks ago, I was a part of <a href="https://hackthenorth.com/">Hack the North’s</a> team as a designer for Battle of the Hacks at Andreessen Horowitz’s HQ. It’s an exclusive 24 hour hackathon for hackathon organizers, with a grand prize of $25,000 for the team’s hackathon. Hack the North won last year with <a href="http://devpost.com/software/moocast">an Android screen streaming app.</a></p><p>This year, we wanted to try something even crazier. We ended up bringing <a href="http://www.theverge.com/2016/5/18/11705918/google-instant-apps-android-hands-on-video">Instant Apps</a> to iOS. It was <strong>a ton of fun </strong>and to our surprise, <strong>it worked. </strong>By the end of the hackathon, our end result could open native iOS apps without installation. We called it OneApp.</p><h4>What is Instant Apps and why do we need it?</h4><blockquote>We’re evolving Android Apps to run instantly without installation. <br> — Ellie Powers, Product Manager at Google</blockquote><p>During Google I/O 2016, Instant Apps was shown to the world. It’s a new concept allowing your device to not have to install any apps. Google is bringing this to Android phones and showed everyone how useful this could be.</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2F9oaaecd7NpI%3Ffeature%3Doembed&amp;url=http%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3D9oaaecd7NpI&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2F9oaaecd7NpI%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/5899730a0c812880092c7441cb740078/href">https://medium.com/media/5899730a0c812880092c7441cb740078/href</a></iframe><p>The current state of apps is that all native apps have to be installed on your devices before they can be used. Native apps cannot be opened directly.</p><p>Similar to how people prefer streaming instead of downloading a movie, Instant Apps allows you to use the app without installing the app. It gets downloaded and lives in your device’s memory until you’re done using it.</p><h3>1. Defining a Problem to Solve</h3><p><em>Every idea starts with a problem. What are you trying to solve? Is the problem worth solving? From the initial problem, it allows you to get the ball rolling and come up with different ideas for your hackathon.</em></p><p>—</p><p>With the introduction of Instant Apps, the experience for using apps is going to be completely redefined. Constantly having to install apps to use them on infrequent occasions adds friction and underutilizes our phone’s limited storage space.</p><p>Google will be releasing this for Android developers and users soon. Since there is currently no way to open iOS apps without installation, we decided to build that experience during the hackathon.</p><p>Due to the time constraints of a hackathon, we had to narrow the scope our hack. In order to do that, we asked ourselves:</p><p><strong>What could improve your experience with your phone if you had access to all existing apps without installation?</strong></p><h3>2. Brainstorming</h3><p><em>After defining the problem, it’s important to create a list of possible solutions. By the end of a hackathon, you want to show that your hack is useful and can add value to its users.</em></p><p>—</p><p>After establishing the constraints, we needed to define how users would use our app. Having access to all apps on your phone opens a whole new world of ideas and experiences. Our team came up with several new use cases and thought through old ones, such as:</p><ul><li>Directly opening a native app from the browser</li><li>Play a demo of a game like Flappy Bird</li><li>Open an app from a QR code</li><li>Receive push notifications from apps you don’t have installed</li></ul><p>We had many more ideas for our app. Given that we had 24 hours, we scoped the project to the best ones.</p><h3>3. Designing the Best Use Cases</h3><p><em>Be selective. From your list of ideas, select the best ones. Keep in mind that you will only have 24 hours for your hackathon. For your top ideas, start asking yourself how these solve your initial problem.</em></p><p>—</p><h4>Browsing Apps</h4><p>The first was being able to use any native app without installing it. When a user needs an app, OneApp would allow users to browse uninstalled apps that they can open directly. Users can search for an app similar to how they’d search for it in their iOS home screen and open it directly.</p><h4>Context Aware Apps</h4><p>We introduced context(time, location, etc…) triggers to prompt users to use an app. This becomes especially useful for one-off apps. For example, event apps(Facebook F8, WWDC, etc…), airport apps, restaurant apps, etc. Each of these require specific contexts (time, location) for the app to be used.</p><p>For example, if your location says you’re at the airport, it’d ask you if you wanted to check in.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/225/1*tMC4yXGh7wp926yYbVVviw.png" /></figure><p><strong>We want to invite users to use an instant app when they need it the most.</strong></p><p>By prompting the user to open the app when they needed it, we hoped to:</p><ol><li>Increase the app’s visibility, allowing the user to know its existence. Without surfacing apps through prompt, they’re buried as one of thousands of apps on the App Store.</li><li>Allow I the user to become more productive. When in a certain context, there’s a high likelihood that a user has a need to be met. Providing that initial prompt allows users to accomplish that need.</li></ol><h4>Deep Linking Apps</h4><p>One of the most powerful features on mobile apps is the ability to deep link into other apps. We use it almost every day when we use our mobile web browser and find ourself opening an installed app.</p><p>At Google I/O, Ellie Powers presented the situation where a friend messages you a Buzzfeed URL. After clicking on it, your phone would open the native Buzzfeed app.</p><p>Our app would become a middle man between the mobile web browser and open the intended native app directly.</p><p><strong>From a user’s perspective, we’d be recreating Instant Apps on iOS.</strong></p><h3>4. Building the Best Use Cases</h3><p><em>This is the step where all of your technical skills come into play. When working in teams, it’s important to flush out design and technical details on how you want to implement the solution to the initial problem.</em></p><p>—</p><p>After establishing the task list for what needed to be done, we were ready for the hackathon.</p><h4>Implementation</h4><p>Currently, opening an iOS app requires the app to be installed. If we wanted to be able to open an app that isn’t installed, we needed something that could run within an iOS app. That’s why we opted for React Native. React Native apps can be interpreted within the iOS app during runtime because it is based on JavaScript.</p><p>Similar to how Google has servers set up to serve Android apps to Android devices, we set up our own servers to store and serve our own apps. When a user wanted to use an app, the app would request for the app from a server with an app ID. The server responds with the source code of the requested app. OneApp would then <a href="https://facebook.github.io/react-native/blog/2016/03/24/introducing-hot-reloading.html">hot reload</a> the requested app</p><h4>Product Design</h4><p><em>It’s important to ask ourselves how to design a solution for those 3 use cases. We want to keep in mind how a user should feel while using our app, and if we’re able to create a good experience that solves the problem at hand.</em></p><p>—</p><p>We wanted OneApp to become the invisible middle man between the user and the apps they wanted to use. In order to allow users to find the app they need, we grouped all apps into 3 levels of visibility: push, recommended and browse.</p><p><strong>Push-level </strong>apps get surfaced in two ways: through push notifications or through opening the app when the user’s context meets specific requirements. For example, you are at the Moscone Center during WWDC. These are brought to your attention if we’re over 90% sure you want to use this app.</p><p><strong>Recommended-level</strong> apps are more loose than push in terms of requirements. These apps depend on some context, or how frequent you use the app.</p><p><strong>Browse-level</strong> apps are all of the other apps. You may have never used them, you may use them frequently, but you aren’t in a context where we want to make them visible to you.</p><p>Depending on the context, the app gets opened in different ways.</p><ul><li>If it was through a push notification, we open the app directly because we know the user was interested in what the app had to offer.</li><li>If the app is in a very specific context and the user opens the app from the home screen, we greet them and ask if they would like to use the most recommended app.</li><li>In most cases where the user opens the app directly, we provide app recommendations as well as the ability to browse all app. This flow is similar to an App Store, but allows a user to immediately open an app versus having install as the only option.</li></ul><h4>Visual Design</h4><p>Good visual design is a key part of many winning hackathon projects. We were in an environment where we had to iterate fast and allow developers to quickly pull assets from the designer’s work.</p><p>We chose to use <a href="http://figma.com/">Figma</a> to design our UI. Since Figma lives in the web, it allowed developers on our team to explore and export all assets through a simple URL.</p><p>—</p><p>The first thing we designed was what our landing page would look like. Since some of our app’s functionality depends on context, we have different UI’s depending on the user’s context.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*i8v4IiugfHxhw1ycmHsfFw.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*eAIAxhxS8H74t1S2_Ue4Ag.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*_2UR12yHflFkw6HbeGDw2w.png" /></figure><p>One of the best use cases are for airport checkin. We designed some mocks for our developers to prototype for our demo.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/281/1*1zICjbxSA8wtm3LJpIivag.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/281/1*_YFT24sNAukKNhIqFT6VNA.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/281/1*vtEtIb80QzuFQR5lEYRZ3w.png" /><figcaption>Mockups for a demo checkin app</figcaption></figure><p>Maybe you want to check out the flights dashboard when you’re at the airport!</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/281/1*JABGYr-parV-qE9nTL5qJQ.png" /><figcaption>Mockup for flights dashboard departing from SFO</figcaption></figure><p>Here are some of the instant apps that can be recommended to our users</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Pc1p4wy90252f3H37q10Tw.png" /><figcaption>Mockups for recommended apps (including MooCast, our winning app from last year)</figcaption></figure><h3>5. Demo</h3><p>We had 5 minutes to present our hack. We split our presentation into three parts:</p><ol><li><strong>Problem:</strong> Why did we build this app, and what we’re trying to solve.</li><li><strong>Demo: </strong>What we managed to implement within 24 hours.</li><li><strong>Technology: </strong>What we used to build this and what challenges we faced</li><li><strong>Potential: </strong>What is the potential for this app</li></ol><p>Skip to 42:20 to see the demo for our app.</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.facebook.com%2Fvideo%2Fembed%3Fvideo_id%3D1151615134889163&amp;url=https%3A%2F%2Fwww.facebook.com%2FMajorLeagueHacking%2Fvideos%2F1151615134889163%2F&amp;image=https%3A%2F%2Fscontent.xx.fbcdn.net%2Fv%2Ft15.0-10%2Fp160x160%2F13481666_1151653224885354_1893804828_n.jpg%3Foh%3D1eacc434c1bb57eab0c78fa437bc331b%26oe%3D57ECE0DF&amp;key=d04bfffea46d4aeda930ec88cc64b87c&amp;type=text%2Fhtml&amp;schema=facebook" width="560" height="315" frameborder="0" scrolling="no"><a href="https://medium.com/media/6d86430c8e90724291f7333f6278ddd7/href">https://medium.com/media/6d86430c8e90724291f7333f6278ddd7/href</a></iframe><h3>Takeaways</h3><p>Hackathons are all about learning, having fun and feeling the satisfaction of making something awesome. For us, we were excited to turn an idea into a reality, and see how instant apps would feel on iOS.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*tHZ12e10T-GCCwNVFviTAA.png" /><figcaption>Team Hack the North after 24 hours of hacking(Michael Yoo, Zheng Tao, Kartik Talwar, Joey Pereira and myself)!</figcaption></figure><p>—</p><p><em>If you are also interested in building ideas, designing solutions to problems or learning about technology and hacking, we want you to come to Hack the North. </em><a href="https://hackthenorth.com/"><em>Apply now</em></a><em>!</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=eb5b2333c9c2" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>