<?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 Rachael Olaniyan on Medium]]></title>
        <description><![CDATA[Stories by Rachael Olaniyan on Medium]]></description>
        <link>https://medium.com/@rachaelo?source=rss-f917ac799823------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*6Da9luNkZwlZ9wHwIoeUqg.jpeg</url>
            <title>Stories by Rachael Olaniyan on Medium</title>
            <link>https://medium.com/@rachaelo?source=rss-f917ac799823------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sun, 22 Nov 2026 14:59:23 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@rachaelo/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[PRD: Cancellation Waitlist for Specialist Appointments (Ontario Pilot)]]></title>
            <link>https://rachaelo.medium.com/prd-cancellation-waitlist-for-specialist-appointments-ontario-pilot-596cc1be2929?source=rss-f917ac799823------2</link>
            <guid isPermaLink="false">https://medium.com/p/596cc1be2929</guid>
            <category><![CDATA[healthcare]]></category>
            <category><![CDATA[healthcare-technology]]></category>
            <category><![CDATA[product-management]]></category>
            <category><![CDATA[product-requirements]]></category>
            <dc:creator><![CDATA[Rachael Olaniyan]]></dc:creator>
            <pubDate>Sat, 14 Nov 2026 21:06:59 GMT</pubDate>
            <atom:updated>2026-11-14T21:06:59.179Z</atom:updated>
            <content:encoded><![CDATA[<p><em>A concept PRD that goes with my </em><a href="https://rachaelo.medium.com/i-gave-up-trying-to-book-a-specialist-in-2020-heres-what-i-d-build-now-as-a-data-product-manager-ed6abe916e55?sharedUserId=rachaelo"><em>Medium article</em></a><em>: It describes a product I haven’t built.</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*vTsPVb01xMXDrruGOJ7fyw@2x.png" /><figcaption>When a spot opens up, the next person on the waitlist gets 24 hours to accept. Concept mockup.</figcaption></figure><h3>1. Problem</h3><p>Patients in Ontario who need a specialist often phone clinics one at a time and hear waits of months. We believe cancelled appointments often go unfilled while patients keep waiting, but this needs to be confirmed with clinics before anything is built.</p><h3>2. Users</h3><ul><li><strong>Primary:</strong> a patient who needs a dermatology visit, already has a referral or sees a service that doesn’t require one, and would take a short-notice slot.</li><li><strong>Secondary:</strong> a clinic administrator or doctor with a cancelled slot to fill.</li></ul><h3>3. Hypothesis</h3><p>If patients can see open and cancelled slots near them and have 24 hours to accept an offer, more cancelled slots will be used and fewer patients will give up.</p><h3>4. Goals and non-goals</h3><p><strong>Goals</strong></p><ul><li>Let a patient find the next available slot in their area, and join a waitlist when there is none.</li><li>Offer cancelled slots to waitlisted patients quickly and fairly.</li><li>Give clinics a simple way to publish availability and fill gaps.</li></ul><p><strong>Not in the pilot:</strong> medical advice or triage, live ER wait times, insurance on file, health card collection, reviews, payments, and the AI booking assistant.</p><h3>5. Pilot scope</h3><p>Dermatology, one region, and about five willing clinics. Patients and clinics are onboarded by hand at first.</p><h3>6. Requirements</h3><p><strong>Must have</strong></p><ul><li>Sign in as a patient or a clinic (both)</li><li>Search by care type, area, and week, with next available times and a next-city fallback (patient)</li><li>Join a waitlist with a date window (patient)</li><li>Notify by push or text when a slot opens (patient)</li><li>Accept or decline within the hold window (patient)</li><li>Publish open slots and mark cancellations (clinic)</li><li>Set each service’s referral rule (clinic)</li><li>Offer the slot to the next person when a hold expires or is declined (system)</li></ul><p><strong>Should have</strong></p><ul><li>Booking confirmation and reminders (both)</li></ul><p><strong>Could have</strong></p><ul><li>Referral code entry, see section 7 (patient)</li><li>See waitlist demand by week (clinic)</li></ul><p><strong>Hold window:</strong> 24 hours by default, shortened when the appointment is close so slots don’t go stale. The exact cutoff should be set with clinics.</p><h3>7. Referral handling (hypothesis)</h3><p>Many publicly covered specialist visits need a family doctor’s referral. Three options, in the order I’d try them:</p><ol><li><strong>Services with no referral needed.</strong> The pilot starts here, so referrals don’t block anything.</li><li><strong>Patient confirms a referral exists.</strong> The clinic verifies it before the visit.</li><li><strong>Referral code.</strong> The doctor’s system issues a one-time code that points to the referral, expires after one use, is tied to the patient, carries the urgency the doctor set, and holds no medical details. This depends on connecting to Ontario’s eReferral system, and I haven’t verified that it is feasible.</li></ol><h3>8. Privacy and safety</h3><ul><li>Collect the minimum personal health information, and get a privacy and legal review (PHIPA) before using real patient data.</li><li>Give no medical advice. If a patient describes something urgent, point them to emergency care.</li><li>Make the offer order visible, for example by time joined adjusted for referral priority, so the process is fair.</li></ul><h3>9. Success metrics</h3><ul><li><strong>Time to first available appointment:</strong> compared with calling clinics yourself.</li><li><strong>Cancelled slots refilled within 24 hours:</strong> refilled slots divided by cancelled slots.</li><li><strong>Waitlist acceptance rate:</strong> accepted offers divided by offers sent.</li><li><strong>Search abandonment:</strong> searches that end with no booking or waitlist join.</li><li><strong>Clinic availability updates:</strong> updates per clinic per week.</li></ul><p>Targets are set after discovery, once real cancellation and response data exist.</p><h3>10. Riskiest assumptions and how to test them</h3><ul><li><strong>Clinics will keep availability current.</strong> Test: pilot with about five clinics and track weekly updates. I’d rethink if updates stop after the first two weeks.</li><li><strong>Cancelled slots are frequent enough to matter.</strong> Test: ask clinics for four weeks of cancellation counts. I’d rethink if there are very few cancellations per clinic.</li><li><strong>Patients respond within the hold window.</strong> Test: a text-only trial with about 20 patients. I’d rethink if most offers expire unanswered.</li><li><strong>Referral rules are manageable.</strong> Test: interview three clinics. I’d rethink if the rules vary too much to encode.</li></ul><h3>11. Phases</h3><ol><li><strong>Discovery (about 2 weeks):</strong> interview clinics and patients, and collect cancellation counts.</li><li><strong>Concierge test (about 2 to 4 weeks):</strong> run the waitlist by hand with a spreadsheet and text messages before building software.</li><li><strong>Prototype and pilot:</strong> build the screens, then run the pilot with the first clinics.</li><li><strong>Later:</strong> the AI assistant, referral codes, and more specialties.</li></ol><h3>12. Open questions</h3><ul><li>Which region and clinics give the best chance of a willing first partner?</li><li>Who sets the offer order when two patients are waiting for the same slot?</li><li>What is the lowest-effort way for a clinic to publish availability?</li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=596cc1be2929" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[I Gave Up Trying to Book a Specialist in 2020. Here’s What I’d Build Now as a Data Product Manager.]]></title>
            <link>https://rachaelo.medium.com/i-gave-up-trying-to-book-a-specialist-in-2020-heres-what-i-d-build-now-as-a-data-product-manager-ed6abe916e55?source=rss-f917ac799823------2</link>
            <guid isPermaLink="false">https://medium.com/p/ed6abe916e55</guid>
            <category><![CDATA[digital-health]]></category>
            <category><![CDATA[canada]]></category>
            <category><![CDATA[healthcare-technology]]></category>
            <category><![CDATA[healthcare]]></category>
            <category><![CDATA[data-product-manager]]></category>
            <dc:creator><![CDATA[Rachael Olaniyan]]></dc:creator>
            <pubDate>Sat, 14 Nov 2026 20:59:01 GMT</pubDate>
            <atom:updated>2026-11-14T21:07:50.056Z</atom:updated>
            <content:encoded><![CDATA[<p><em>A story about a keloid, a lot of phone calls, and a problem that turned out to be about data.</em></p><p>September 2020. COVID-19 was about to take over the world, and I had a large keloid on my skin that I couldn’t ignore. It was uncomfortable, and it didn’t feel like something that could wait.</p><p>So I did what most of us do. Visited a doctor, got referred and was asked to wait a year. I googled dermatologists near me. Then I copied down names, addresses, and phone numbers, one tab at a time. Then I started calling.</p><p>Every clinic had its own answer, and the answers all sounded the same: six months, maybe a year.</p><p>Somewhere along the way I stopped. I didn’t stop because I’d found a solution. I stopped because it was tedious, I was tired, and my issue needed attention now, not next year. I went back to at-home remedies and told myself I’d deal with it later.</p><p>I didn’t think of it as a product problem back then. I was just a patient. But the experience stuck with me, and after moving from data analytics into data product management, I can’t stop seeing it.</p><h3>It wasn’t one problem. It was three.</h3><p>Looking back, what I went through had three parts:</p><ol><li><strong>Finding.</strong> There was no single place to see who treats what, near me.</li><li><strong>Asking.</strong> Every clinic answered by phone, one at a time, and I had to repeat myself each time.</li><li><strong>Waiting.</strong> Nobody could tell me if a spot might open up sooner.</li></ol><p>I’ll be honest about something up front. The third one, the long wait itself, is mostly a capacity problem. There aren’t enough appointments to go around, and no app creates a dermatologist. But the first two are information problems, and information problems are what I work on.</p><h3>You’re not imagining it: what the numbers say</h3><p>The jokes about being healed before you get an appointment exist for a reason. Here’s what I found when I went looking.</p><ul><li>The Fraser Institute’s 2025 <em>Waiting Your Turn</em> survey found a <a href="https://www.fraserinstitute.org/STUDIES/WAITING-YOUR-TURN-WAIT-TIMES-FOR-HEALTH-CARE-IN-CANADA-2025">national median of 28.6 weeks</a> from a GP’s referral to treatment, the second longest since the survey began in 1993.</li><li>Ontario had the shortest median of any province at 19.2 weeks, down from 23.6 the year before. The wait just to see a specialist was <a href="https://www.todocanada.ca/health-care-wait-times-in-canada-2025-how-long-you-need-to-wait-for-medical-treatment/">about 10.7 weeks</a>, tied with Quebec as the shortest.</li><li>Ontario’s government says the number of people without a regular health care provider has <a href="https://unpublished.ca/news-feed-item/2026-01-12/ontario-touts-drop-in-number-of-residents-lacking-primary-care-providers">dropped to 1.98 million</a>, while the Ontario Medical Association has <a href="https://www.ctvnews.ca/toronto/article/jones-says-ontario-on-track-to-meet-goals-for-connecting-more-residents-with-primary-care/">maintained its estimate of 2.5 million</a>. The two sides measure it differently.</li><li>On ER wait times, <a href="https://erstat.ca/transparency">Ontario Health doesn’t run a centralized system</a>. Individual hospitals publish their own data on different platforms.</li></ul><p>A few caveats, because a data person should give them. The Fraser numbers come from a physician survey with a <a href="https://www.todocanada.ca/health-care-wait-times-in-canada-2025-how-long-you-need-to-wait-for-medical-treatment/">13.1% response rate</a>, and the institute is a policy think tank with its own perspective. Medians also hide the long tail, and my own phone calls were the tail. So treat the numbers as a signal, not a verdict. The signal is still clear.</p><h3>What a data product manager sees</h3><p>Much of my work is helping teams build a single source of truth, one trusted place where data lives so people stop chasing it across systems. When I look at booking a specialist, I see the opposite: availability data scattered across hundreds of separate clinic systems and phone lines, with the patient acting as the human integration layer.</p><p>The data itself isn’t exotic. At its core, you’d need:</p><ul><li><strong>Providers:</strong> who they are, their specialty, location, and what they treat.</li><li><strong>Slots:</strong> open and cancelled appointments with a date and time.</li><li><strong>Rules:</strong> whether a referral is needed, whether they’re accepting new patients, whether the visit is virtual or in person.</li><li><strong>Patients:</strong> what they’re looking for, their area, and their flexibility.</li></ul><p>The product, then, is about matching and notification. Who needs a slot, and who just freed one up?</p><h3>The product I’d build</h3><p>This is a concept. I haven’t built it, and I want to be clear about that. Here’s the experience I imagine.</p><ul><li><strong>One front door, two sides.</strong> You sign in as a patient or as a doctor. Patients are there to find and book care. Doctors and clinics are there to publish availability and the services they offer.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/390/1*xbXx4ZlfHj0lxjXP1Gk1Wg@2x.png" /><figcaption><em>[Figure 1: Sign-in screen with patient and doctor options]</em></figcaption></figure><ul><li><strong>Search that works the way you think.</strong> You tell the app what kind of care you need, when you’d like to be seen, and where. It shows who has openings in that window near you. If nobody nearby does, it shows the next city over, so you decide whether the drive is worth it instead of calling to find out.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/390/1*GnsQTxt-cD-De3HLptw_ew@2x.png" /><figcaption><em>[Figure 2: Search results with the next available times and a nearby-city fallback.]</em></figcaption></figure><ul><li><strong>A cancellation waitlist.</strong> This is the idea I care about most. If your preferred doctor has nothing in your window, you join a waitlist. When someone cancels, you’re notified and given 24 hours to accept. If you don’t, the slot goes to the next person, so nothing sits empty. One design detail matters here: as the appointment gets closer, that 24-hour window has to shrink, so a slot tomorrow morning doesn’t wait a full day.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/390/1*P7THWsMqhK614rJ2hbqGrg@2x.png" /><figcaption><em>[Figure 3: Waitlist confirmation and the “a spot opened up” notification with a 24-hour countdown.]</em></figcaption></figure><ul><li><strong>A referral code that follows you.</strong> This idea comes from my own experience of being referred to a dermatologist. When your doctor refers you, their system also gives you a short, one-time code. You enter it in the app, and the specialist’s clinic already sees the referral, so nobody has to ask who sent you or why. The code would only point to the referral and never carry medical details, and it would expire and work once. It would also be tied to you, so it can’t be passed around. Specialists often triage referrals by urgency, so the code should carry the priority your doctor set, and the app shouldn’t let anyone jump the queue. Done right, a referred patient could book into open slots instead of waiting for a clinic to call. It would depend on connecting to <a href="https://nygh.on.ca/ereferral">Ontario’s existing eReferral system</a>, and I haven’t verified that this is feasible, so treat it as a hypothesis to test with a clinic, not a promise.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/390/1*tc_lR9g2UHuYXO1XvQCeJA@2x.png" /><figcaption>[Figure 4: Enter the one-time code from your doctor, and the clinic already sees your referral.]</figcaption></figure><ul><li><strong>An assistant for the tedious part.</strong> Instead of filling in filters, you could just say: “I’d like to see a dermatologist the week of November 23.” The AI chat assistant looks at real availability, shows you the options, and books the one you pick after you confirm. It would only work from real calendar data and never invent an opening. It wouldn’t give medical advice, and if you describe something that sounds urgent, it would point you to emergency care or a nurse line rather than a booking.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/390/1*qx3kC-JP81jtOZKtqGFMCA@2x.png" /><figcaption><em>[Figure 4: AI Chat conversation booking a dermatologist for a specific week.]</em></figcaption></figure><h3>Who it’s for, and what I’d build first</h3><p><strong>The user:</strong> a patient in Ontario who needs a specialist or allied-health visit, either already has a referral or sees a specialty that doesn’t need one, and would take a short-notice slot. The second user is the clinic with a cancelled appointment it wants to fill.</p><p><strong>The bet:</strong> if patients can see cancelled slots near them and take one within 24 hours, more of those slots get used and fewer people give up after a few phone calls.</p><p><strong>Build first:</strong></p><ul><li>One specialty (dermatology), one region, and a handful of willing clinics.</li><li>Patient search, a simple way for clinics to enter availability, and the cancellation waitlist with the 24-hour hold.</li></ul><p><strong>Don’t build yet:</strong> the AI assistant, live ER wait times, insurance on file, reviews, and other specialties. Each is useful, but none is needed to test the bet.</p><p><strong>Riskiest assumptions to test first:</strong></p><ul><li>Clinics will keep availability current.</li><li>Patients will respond inside the hold window.</li><li>Cancelled slots come up often enough to matter.</li></ul><h3>What I’d be careful about</h3><p>I’d rather name the hard parts myself than have someone point them out.</p><ul><li><strong>Referrals.</strong> Many publicly covered specialist visits start with a family doctor’s referral. <a href="https://www.getmaple.ca/?p=21857">Some specialties, like dermatology, may not need one</a>, so a pilot there makes sense, but the app has to know each specialty’s rules. Ontario also already runs a provincial eReferral program that most clinicians can use from their medical records system, so I wouldn’t reinvent referrals. I’d build on it, as in the referral code idea above.</li><li><strong>Getting the calendars.</strong> Clinics run on different systems. <a href="https://ontariohealth.ca/online-appointment-booking">Ontario Health already recommends online booking</a> and has published a standard for it, which helps, but the integration work is the hard part, not the app screens.</li><li><strong>Privacy.</strong> Health information is sensitive and regulated. Storing health card details raises the bar for security and compliance.</li><li><strong>Fairness.</strong> Reviews of clinics need guardrails so they don’t become a way to unfairly attack individual doctors.</li><li><strong>I’m not the first to try.</strong> Services like <a href="https://www.getmaple.ca/regions/ontario/">Maple</a> offer virtual specialist visits, and platforms like <a href="https://cortico.health/specialties/ophthalmologist/ontario/?view=map">Cortico</a> let people book online with clinics. The question isn’t whether booking tools exist. It’s whether the waitlist and the single view across clinics are done well enough to change the experience.</li><li><strong>It won’t fix capacity.</strong> A tool that shows openings can only redistribute the appointments that exist. What it can do is make sure cancelled slots get used and fewer people give up.</li></ul><h3>How I’d know it works</h3><p>I wouldn’t start with seven features. I’d start with one specialty, one region, and a handful of willing clinics. Then I’d track:</p><ul><li><strong>Time to first available appointment</strong> compared with calling around.</li><li><strong>Cancelled slots refilled</strong> within 24 hours.</li><li><strong>Waitlist acceptance rate</strong> within the 24-hour window.</li><li><strong>No-show rate</strong> for clinics using the waitlist.</li><li><strong>Drop-off:</strong> how many people start a search and abandon it, which is the 2020 me.</li></ul><h3>Why I’m writing this</h3><p>I don’t know if an app would have changed my outcome in September 2020. What I do know is that giving up shouldn’t be the default response to an overloaded system. Behind every “six months to a year” is a person who has to decide whether to keep calling.</p><p>If you’ve been through this too, in Ontario or anywhere else, I’d like to hear how you handled it and what would have helped most.</p><p><strong>Read the full product spec (PRD) on </strong><a href="https://rachaelo.medium.com/prd-cancellation-waitlist-for-specialist-appointments-ontario-pilot-596cc1be2929?sharedUserId=rachaelo"><strong>Medium</strong></a></p><h3>Sources and further reading</h3><ul><li>Fraser Institute, <a href="https://www.fraserinstitute.org/STUDIES/WAITING-YOUR-TURN-WAIT-TIMES-FOR-HEALTH-CARE-IN-CANADA-2025"><em>Waiting Your Turn: Wait Times for Health Care in Canada, 2025</em></a></li><li>TodoCanada, <a href="https://www.todocanada.ca/health-care-wait-times-in-canada-2025-how-long-you-need-to-wait-for-medical-treatment/">Health care wait times in Canada 2025</a></li><li>Unpublished, <a href="https://unpublished.ca/news-feed-item/2026-01-12/ontario-touts-drop-in-number-of-residents-lacking-primary-care-providers">Ontario touts drop in number of residents lacking primary-care providers</a></li><li>CTV News, <a href="https://www.ctvnews.ca/toronto/article/jones-says-ontario-on-track-to-meet-goals-for-connecting-more-residents-with-primary-care/">Jones says Ontario on track to meet goals for connecting more residents with primary care</a></li><li>ERstat, <a href="https://erstat.ca/transparency">Canadian ER Data Transparency Report Card</a></li><li>Ontario Health, <a href="https://ontariohealth.ca/online-appointment-booking">Online Appointment Booking</a></li><li>North York General Hospital, <a href="https://nygh.on.ca/ereferral">Ocean eReferral overview</a></li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=ed6abe916e55" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Disney+ Has the Best Content in the World. Its Platform Is Holding It Back.]]></title>
            <link>https://rachaelo.medium.com/the-content-is-magic-the-infrastructure-is-friction-a-strategic-audit-of-disney-5c97cd6986a1?source=rss-f917ac799823------2</link>
            <guid isPermaLink="false">https://medium.com/p/5c97cd6986a1</guid>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[user-experience]]></category>
            <category><![CDATA[product-management]]></category>
            <category><![CDATA[product-strategy]]></category>
            <dc:creator><![CDATA[Rachael Olaniyan]]></dc:creator>
            <pubDate>Mon, 16 Feb 2026 20:39:20 GMT</pubDate>
            <atom:updated>2026-06-25T21:09:31.362Z</atom:updated>
            <content:encoded><![CDATA[<p>Streaming is no longer a content war. It stopped being that a few years ago.</p><p>Every major studio now has a platform. The race to build the biggest content library has reached a point where no single catalogue can sustainably hold an audience on its own. What separates the winners going forward is infrastructure. Specifically, the ability to use AI to understand users in real time, reduce friction between browsing and watching, and build a platform that gets smarter the more people use it.</p><p>Netflix figured this out early. Their recommendation engine drives over 80% of what subscribers watch, and the company estimates that AI-powered personalization saves them over $1 billion annually in retention revenue alone. That is not a side feature. That is the product.</p><p>Disney+ is in a different position. It has arguably the most powerful content library in the world and 125 million subscribers as of early 2025. But the platform underneath that library was built as a content vault, not an intelligent delivery engine. And that gap is starting to cost them.</p><p>In the final quarter of 2024, Disney+ lost 700,000 subscribers. Downloads in the US hit their lowest quarterly figure since the platform launched. Disney’s new CEO Josh D’Amaro has publicly committed to AI as a core strategic pillar across five business areas, including personalization and monetization on Disney+. The intent is clearly there. The execution is what matters now.</p><p>So what would it actually take to rebuild Disney+ as an AI-first platform? Here is how what I think about it;</p><h3>The Core Problem: The Platform Does Not Learn From You</h3><p>Before jumping to solutions, it is worth being precise about the problem. Disney+ is not a bad product. The interface is clean, the content library is deep, and the streaming quality is reliable. The issue is that the platform is largely static. It does not adapt in real time to how you actually watch.</p><p>It does not know the difference between a user who opened the app with thirty minutes to spare and one who just settled in for a three-hour family movie night. It serves content. It does not learn from how you receive it.</p><p>Netflix, by comparison, processes hundreds of billions of user events every single day. Pauses, rewinds, what you stopped watching twelve minutes in, what time of day you tend to watch, which thumbnail made you click. That behavioural data feeds back into a recommendation engine that refines itself with every session. The result is a platform that feels like it knows you, because it genuinely does.</p><p>Disney+ generates the same signals from tens of millions of active users every day. What it lacks is the infrastructure to turn those signals into a smarter experience. That is the gap, and it is a platform problem before it is anything else.</p><h3>What the AI Opportunity Actually Looks Like</h3><p>Disney’s new leadership has outlined AI investment across content creation, workforce productivity, guest experiences, and streaming personalization. The streaming piece is where the user impact will be most immediate, and also where the competitive bar is highest, because Netflix has been building for over twenty years.</p><p>There are three areas where I think the investment matters most.</p><p><strong>1. Discovery that reduces how long it takes to find something worth watching</strong></p><p>The average streaming user spends more time deciding what to watch than actually watching. Research from Limelight Networks found that around 20% of viewers abandon a streaming session entirely when they cannot find something within a few minutes. On a platform with over 13,000 titles, the cost of a weak discovery experience is enormous.</p><p>The current Disney+ home screen puts the entire decision on the user. You scroll, you hover, you read a generic description, and then you decide. For a library that size, the platform should be doing a lot more of that work.</p><p>The AI opportunity here is personalized, context-aware discovery. Not just “because you watched Moana” logic, but a system that reads your current session. Are you browsing alone late at night? Are you looking for something to put on for young children? Did you just finish a six-episode series? A well-trained model should be picking up on those signals and surfacing content accordingly, not just pulling from a static watch history.</p><p>The metric that matters here is Time-to-Play: how long between opening the app and actually pressing play. If the discovery layer is working, that number falls. If it is not, you have added friction to the home screen and made things worse.</p><p><strong>2. Autoplay logic that reflects whether someone is actually still watching</strong></p><p>This is a smaller issue but it reveals a lot about how the platform was built. Disney+’s autoplay timer is fixed. After a set number of episodes, it shuts off and assumes the viewer has left. If you manually resume, the platform does not register that. It will shut off again on the same schedule.</p><p>Netflix’s “Are you still watching?” prompt works differently. It responds to real engagement signals like volume changes, interaction patterns, and device activity. It is not a timer running in the background. It is a model making an inference.</p><p>The reason this matters goes beyond convenience. Every unintended session exit is a behavioural signal the platform fails to capture. Every time a user has to manually resume after an unnecessary shutdown, that is friction that quietly chips away at trust. A 2025 industry study found that reducing unintended session exits by even 5% had a measurable impact on monthly active usage and subscription renewal rates.</p><p>A smarter engagement model would track active session signals, reset when a user resumes manually, and use those patterns to make better content recommendations over time. That is the compounding effect: every session should leave the next one a little sharper.</p><p><strong>3. Global infrastructure that actually matches global content ambitions</strong></p><p>This is the angle that gets talked about the least, but it represents the largest long-term opportunity.</p><p>Disney+ is available in 150 countries and produces content specifically for global audiences. Iwájú, their animated series set in a futuristic Lagos, is a clear example. It is a high-budget, culturally specific production made for a Nigerian audience, on a platform that is not officially available in Nigeria.</p><p>That is not a content problem. It is a business model problem.</p><p>Nigeria has over 240 million people, the largest population on the African continent, and a fast-growing mobile internet economy. Netflix recognised this years ago. They responded with Nollywood partnerships, mobile-only pricing, and payment options that reflect how people in those markets actually transact. The result is that by the time Disney produces content for that audience, its primary competitor is already there.</p><p>Getting into these markets is not simply a matter of switching on availability in a new country. It requires partnering with local payment providers like PalmPay and OPay. It requires streaming that is built for variable bandwidth conditions, not a reduced version of the North American product. And it requires pricing that reflects local purchasing power, not a converted US subscription price.</p><p>The financial case is straightforward. Disney’s content spend has fallen from $29.9 billion in 2022 to an estimated $25 billion in 2024, while many of the markets that could generate returns on that content remain closed to the platform. That is investment without full revenue recovery. Fixing it is a platform and market entry decision, not a creative one.</p><h3>What Disney Is Getting Right</h3><p>It is worth being fair. Disney’s 2025 numbers are not entirely bad. Direct-to-Consumer operating income reached $1.3 billion for the year, a meaningful turnaround from a $4 billion loss just three years prior. The bundle strategy with Hulu and ESPN+ is working. And CEO Josh D’Amaro has shown real urgency around AI investment, with a publicly stated goal of double-digit streaming growth.</p><p>The early AI applications being explored, including personalized recommendations, vertical video formats, and community-generated experiences, show that the product thinking is moving forward. Disney also made a significant move by licensing its IP to OpenAI for use within Sora, though that partnership subsequently ended. What that episode reveals is that the company is still working out where the boundary is between embracing AI and protecting the brand and creative integrity it has built over a hundred years. That is a legitimate and genuinely difficult tension to navigate.</p><p>Disney is a company that has always cared deeply about how its stories are told and experienced. Allowing an AI layer have an influence over that experience is not a decision to take lightly. But the alternative is operating a platform that stays largely the same while competitors build systems that learn and improve with every single viewing session. That path leads somewhere slow and difficult.</p><h3>The Bottom Line</h3><p>Disney+ has one structural advantage that no competitor can buy: its intellectual property. Star Wars, Marvel, Pixar, the Disney classics. That catalogue is not going anywhere. But in a market where streaming subscriptions are discretionary and a subscriber can cancel in thirty seconds, content is what gets someone to sign up. The platform is what keeps them.</p><p>Right now, the platform is not earning that loyalty at the level the content deserves. The infrastructure is not learning fast enough. The global footprint is not keeping pace with the global content ambitions. And while the AI investment is pointed in the right direction, it is still early compared to where Netflix has been operating for the better part of a decade.</p><p>The gap is closeable. Disney has the data, the audience, the capital, and leadership that has now publicly committed to the work. What it needs is the product discipline to follow through: to stop thinking of Disney+ as a place where content lives and start building it as a platform that gets meaningfully better every time someone presses play.</p><p>The content is already there. The platform has to catch up.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=5c97cd6986a1" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Data is the Architect: Why the Best Product Decisions are Made Before the First Line of Code]]></title>
            <link>https://rachaelo.medium.com/data-is-the-architect-why-the-best-product-decisions-are-made-before-the-first-line-of-code-4fd081aa816f?source=rss-f917ac799823------2</link>
            <guid isPermaLink="false">https://medium.com/p/4fd081aa816f</guid>
            <dc:creator><![CDATA[Rachael Olaniyan]]></dc:creator>
            <pubDate>Fri, 23 Jan 2026 00:03:43 GMT</pubDate>
            <atom:updated>2026-07-28T01:28:44.113Z</atom:updated>
            <content:encoded><![CDATA[<blockquote><strong>Data is the Architect: Why the Best Product Decisions Are Made Before the First Line of Code</strong></blockquote><p>A $0.00 balance on a Thursday afternoon taught me that data isn’t just for measuring results. It’s for uncovering the human anxiety that shapes a growth engine.</p><p>Most people think data’s role starts after a product launches, to check if it worked. I’ve come to believe the opposite: data is the architect of a product, and its most important work happens before a single line of code is written.</p><p>As one of the founding analysts at a high-growth fintech startup, I sat at the intersection of Product, Engineering, and Data from day one. Being in the room while the foundations were built meant I could make sure data was treated as the core driver of what we built, not an afterthought bolted on after launch. One of the products I shaped in that seat was a credit-building tool, and the decisions I made in the discovery phase are still the part of that work I’m most proud of.</p><p><strong>Starting with the mess, not the roadmap</strong></p><p>We didn’t start with “let’s build a credit tool.” We started with the messy reality of our users. I brought my own experience as a newcomer to Canada, the specific hurdle of needing a credit history before you can get almost anything done, into the room alongside a colleague’s account of rebuilding credit after an adverse credit event. Neither story was data yet. They were hypotheses.</p><p>So I went looking for the pattern in our behavioural data, and found one: a segment of highly active users, 30+ transactions a month, whose balance hit $0.00 like clockwork a few days before every payday. On paper, that looks like a cash-flow problem. I made the call that it wasn’t. These users weren’t short on money; they were “credit-thin,” with plenty of activity moving in and out but no vehicle to convert it into savings or credit history. The data showed a correlation between new-to-country status and the absence of traditional credit products, which is exactly the gap I’d lived through myself. At that point I reframed the problem for the team: we didn’t need a feature, we needed a bridge to financial inclusion. That reframe is what got funded, not the original feature request.</p><p><strong>The tradeoff: simple and automated vs. flexible and manual</strong></p><p>Before full-scale development, we used survey and behavioral data to simulate demand, and I ran discovery specifically hunting for hesitation, the small moments where a prototype felt intimidating rather than useful. What we found was that for a newcomer, “credit” isn’t a neutral financial term. It’s a source of anxiety. Session recordings showed people pausing, backing out, re-reading the same screen twice.</p><p>That data put me at odds with part of the team. The engineering-favored path was a manual, configurable subscription model: more flexible, faster to ship, easier to A/B test later. I pushed for something narrower, a single, automated, low-configuration flow, on the bet that removing decisions would matter more to this segment than adding flexibility. I made that case using the hesitation data, not intuition, and it’s the kind of tradeoff a feature list alone would never have surfaced. We went with automated by default.</p><p><strong>What actually happened</strong></p><ul><li><strong>Adoption:</strong> the credit-thin segment signed up and activated at a 30% higher rate than the general user base, a strong signal that the “bridge to financial inclusion” framing, not the original feature request, was the right problem to solve.</li><li><strong>Retention:</strong> users who adopted the tool retained 25% better at 90 days than users who didn’t, evidence that the automated, low-friction flow was easier to stick with, not just easier to build.</li><li><strong>Profitability / LTV:</strong> account profitability for this cohort rose 40% over time, turning a segment we once wrote off as low-cash-flow into some of our most durable accounts.</li><li><strong>Fraud:</strong> the automated model held fraud detection accuracy at 90% or above, matching the manual alternative. Removing friction for users didn’t mean loosening our own guardrails.</li></ul><p>Those numbers moved in the direction the discovery data predicted. That’s the difference between a good story and a good decision.</p><p><strong>The takeaway</strong></p><p>The lesson I keep coming back to: data is most powerful when it validates a human tension, not just when it reports on system performance. When we stopped treating “credit-thin” as a number on a dashboard and started treating it as a real constraint on people’s lives, we built a way in, not just another feature.</p><p>That’s become a framework I return to on every product I touch, not just this one:</p><ol><li><strong>Find the anomaly, not the average.</strong> The $0.00 pattern wasn’t visible in aggregate metrics. It only showed up when I looked at behavior instead of balances.</li><li><strong>Name the tradeoff before you build.</strong> Simple and automated vs. flexible and manual was a real decision with a real cost either way. Make it explicit, make the call, and be ready to defend it with data.</li><li><strong>Let the qualitative signal set the bar, and the quantitative signal confirm it.</strong> Hesitation in a session recording told us where to simplify. Retention and adoption numbers told us whether it worked.</li></ol><p>For me, that’s the real job of a Product Manager: using every tool available, from SQL queries to user empathy, to turn a messy reality into a decision someone’s willing to bet a roadmap on.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=4fd081aa816f" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[From Group Chat to Departure Gate: Why ‘Multiplayer Finance’ is the Next Frontier for Fintech…]]></title>
            <link>https://rachaelo.medium.com/from-group-chat-to-departure-gate-why-multiplayer-finance-is-the-next-frontier-for-fintech-74a8904857cd?source=rss-f917ac799823------2</link>
            <guid isPermaLink="false">https://medium.com/p/74a8904857cd</guid>
            <dc:creator><![CDATA[Rachael Olaniyan]]></dc:creator>
            <pubDate>Wed, 21 Jan 2026 18:06:17 GMT</pubDate>
            <atom:updated>2026-01-21T18:06:17.322Z</atom:updated>
            <content:encoded><![CDATA[<p><strong>From Group Chat to Departure Gate:</strong> Why ‘Multiplayer Finance’ is the Next Frontier for Fintech Growth</p><p>From Group Chat to Departure Gate: Why ‘Multiplayer Finance’ is the Next Frontier for Fintech Growth</p><p>Every financial institution on the planet is fighting a losing battle: High CAC and Low Loyalty. They spend hundreds of dollars to acquire a user, only for them to treat the financial institution as a “utility” , a place to park money until they move it elsewhere.</p><p><strong>The Problem:</strong></p><p>Banks are focused on “Personal Finance.” But life isn’t personal; it’s social. The #1 reason young professionals fail to save isn’t a lack of features, it’s social friction. We’ve all seen the “Bali 2026” group chat die because of Economic Asymmetry. When the planner is ready but the group isn’t, the plan fails, and the money stays stagnant.</p><p><strong>The Solution:</strong></p><p><strong>SquadPact (Multiplayer Banking);</strong> We are turning the “Solo Save Vault” into a <strong><em>Shared Commitment Device</em>.</strong> Imagine Maya. She wants a girls’ trip to Bali. She doesn’t just open a savings account; she starts a <strong>Pact.</strong></p><ol><li><strong>The Viral Spark:</strong> She sends a link to 4 friends. To join the Pact, they <em>must</em> join the bank. This isn’t marketing; it’s <strong>Network Effect acquisition.</strong></li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/696/1*VdpzB6ZrJlrDEAwELI1CnA.png" /></figure><p>2. <strong>The Social Moat:</strong> Every Friday, $50 moves automatically from their chequing account into their group savings (Pact). But the magic is the <strong>Shared Progress Bar.</strong> When Layla sees Maya is at 60%, she feels the social pressure to keep up.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/635/1*J298YKBW7wZan6HO2YC5aQ.png" /></figure><p>3. <strong>The Lock-In:</strong> By “locking” these funds for a social goal, we create <strong>Predictable AUM.</strong> We aren’t just a wallet; we are the engine of their social life.</p><p><strong>The “Billion Dollar” Success Metrics</strong></p><pre><strong>CATEGORY  </strong>  |<strong> KPI                      | THE &quot;SO WHAT?&quot;</strong><br>------------|--------------------------|---------------------------------------------------<br><strong>GROWTH </strong>     | Viral Coefficient (&gt;1.2) | Every 1 &quot;Planner&quot; brings in 4+ new funded accounts.<br><strong>RETENTION</strong>   | The Social Moat          | Users are 5x less likely to churn; friends are watching.<br><strong>REVENUE </strong>    | Goal Completion          | 5% affiliate cut via &quot;One-Click&quot; travel booking.</pre><p>While focusing on the Billion Dollar success metrics, its also important to focus on the risk of this feature.</p><p><strong>The “Guardrail” Metrics (Risk Management)</strong></p><ul><li><strong>Internal Cannibalization Rate:</strong> Are users moving money into SquadPacts that they would have deposited into our high-margin products anyway? We want “new” money, not “reshuffled” money.</li><li><strong>Customer Support Volume per Squad:</strong> If 10% of Squads require manual intervention (e.g., “Liam wants to leave the group and it’s causing an error”), the feature isn’t scalable.</li><li><strong>The “Defect” Rate:</strong> The percentage of automated weekly deposits that fail due to NSF (Non-Sufficient Funds). This helps the bank manage liquidity risk.</li></ul><p><strong>The “Efficiency” Metrics (Unit Economics)</strong></p><ul><li><strong>Conversion to Affiliate Revenue:</strong> What % of users who hit 100% of their goal actually click the “Book Flight” button? This measures the success of the “Departure Gate” transition.</li><li><strong>Incremental AUM Cost:</strong> How much does it cost us in “Praise” notifications and UI maintenance to gain $1 in new deposits compared to a standard marketing campaign?</li><li><strong>Time-to-Fund:</strong> How many days pass between a user receiving a Squad invite and making their first $50 deposit? (Speed = Growth).</li></ul><p><strong>The “Social Health” Metrics (The Moat Quality)</strong></p><ul><li><strong>“Poke” Effectiveness:</strong> What % of users who receive a “Friendly Poke” make a deposit within 24 hours? This proves the “Social Moat” actually works.</li><li><strong>Squad Size Elasticity:</strong> Do groups of 3 save more consistently than groups of 10? Finding the “sweet spot” for group size allows you to optimize the invite flow.</li><li><strong>Referral Velocity:</strong> The speed at which a single “Pact” link spreads through a contact list.</li></ul><p>In summary, <strong>SquadPact</strong> is a multiplayer financial framework that transforms communal goals from group trips, dream homes and weddings to parental gifts, into realized milestones. By replacing “solo saving” with a <strong>Shared Social Ledger</strong>, we solve the problem of financial follow-through via automated contributions and peer accountability. For the user, it’s the social glue that ensures group promises are kept; for the platform, it’s a high-retention growth engine that captures high-intent capital and lowers acquisition costs through built-in network effects.</p><pre>USE CASE        | THE &quot;SOCIAL JOB&quot; (WHY)             | THE BUSINESS VALUE<br>----------------|------------------------------------|-----------------------------------<br>The Big Gift    | Buying a parents&#39; anniversary car  | High Viral Spark (5+ Invites/Pact)<br>The Milestone   | Couples saving for a down payment  | High LTV (Long-term, high AUM)<br>The Safety Net  | Roommates&#39; &quot;Emergency fund.&quot;       | Sticky Retention (Daily Utility)<br>The Experience  | Group trips or Music Festivals.    | Affiliate Revenue (Travel/Tickets)</pre><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=74a8904857cd" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Data Doesn’t Lie: Why “Messy” Data is the Secret Ingredient to Product Strategy]]></title>
            <link>https://rachaelo.medium.com/data-doesnt-lie-why-messy-data-is-the-secret-ingredient-to-product-strategy-8e7bb04bb816?source=rss-f917ac799823------2</link>
            <guid isPermaLink="false">https://medium.com/p/8e7bb04bb816</guid>
            <dc:creator><![CDATA[Rachael Olaniyan]]></dc:creator>
            <pubDate>Wed, 21 Jan 2026 14:22:27 GMT</pubDate>
            <atom:updated>2026-01-21T14:22:27.882Z</atom:updated>
            <content:encoded><![CDATA[<p><strong>Data Doesn’t Lie: </strong>Why “Messy” Data is the Secret Ingredient to Product Strategy</p><p>In every meeting I have sat in, whether it was a high growth fintech startup, a healthtech startup or a massive provincial energy establishment, the same scene always plays out. Brilliant people are sitting around a table or on a zoom call, each of them with a different “gut feeling” about which feature to build next. The VP of finance sees a cost center, the Engineer wondering about a technical debt hurdle while the sales team may see a lost deal.</p><p>In these moments, my job isn’t just to provide a chart. My job is to find the truth.</p><p>Early in my career, I got to realize that most people are afraid of messy data. They see inconsistent user behaviour metrics, inaccurate KPI metrics, fragmented clinical logs, or legacy system exports and they look the other way. They wait for clean data that may never come. But one thing I have learned in my years of experience is that the mess is where gold is buried. If the data is messy, it’s because the human reality behind it is complex. And where there is complexity, there is an opportunity for a product to solve a real problem.</p><p>I remember a specific moment during the peak of the pandemic (Covid-19) while working in a healthtech startup that was responsible for providing Covid19 results(numbers) to the federal and provincial government. We were the middleman between hospitals, essential workplaces, and the federal/provincial government. The data coming in was a disaster; multiple formats, manual entries, and high-stake urgency. Most analysts would have just tried to clean the data for a weekly report.</p><p>Instead, I looked at the mess and saw a product opportunity.</p><p>By analyzing the friction points in how results were submitted, I realized we didn’t just need a better reporting format/system, we needed a submission product. I led the development of a Provider Self-Service Portal that utilized automated field-mapping and error-flagging. We moved away from receiving fragmented spreadsheets to a structured web-interface that guided the user through the submission process. It turned a 24-hour manual verification process into a 5-minute automated upload. The impact this feature also had was eliminating the need for back-and-forth emails to fix “messy” data because the product rejects incorrect data instantly. By the time we launched, we hadn’t just cleaned the data, we had eliminated the administrative burden for thousands of frontline workers.</p><p>That experience solidified my core philosophy: Data doesn’t lie. People can be biased, roadmaps can be political, and gut feelings can be wrong. But if you have the grit to mine the mess, the data will tell you exactly what your product needs to be.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=8e7bb04bb816" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[THE TRANSLATOR: BRIDGING THE GAP BETWEEN ENGINEERING, PRODUCT AND C-SUITE]]></title>
            <link>https://rachaelo.medium.com/the-translator-bridging-the-gap-between-engineering-product-and-c-suite-c5ae7c81bb58?source=rss-f917ac799823------2</link>
            <guid isPermaLink="false">https://medium.com/p/c5ae7c81bb58</guid>
            <dc:creator><![CDATA[Rachael Olaniyan]]></dc:creator>
            <pubDate>Mon, 19 Jan 2026 23:56:37 GMT</pubDate>
            <atom:updated>2026-01-19T23:56:37.118Z</atom:updated>
            <content:encoded><![CDATA[<p>In over 6 years of my career, I have learned that data is a universal language, but it has many different dialects.</p><p>As a data/product leader, my most important role is as a translator. I sit in the middle of a high pressure triangle: the Executives who care about the bottom line, the Engineers who care about system integrity, and the users who just want the product to work.</p><p>When I sit down with a VP of finance or a CEO, I don’t talk about SQL join types or API latency. I talk about revenue risk and market opportunity. When I presented the Grid Stability solution at an energy organization, the story wasn’t about data warehouses, building tables, cleaning data or building pipelines — it was about protecting the organization from the massive financial and reputational cost of an outage. I used data to transform a technical nice-to-have into a business must-have.</p><p>When I turn to the Engineering team, the dialect shifts. Now, the story is about scalability, Caveats and Logic. At the healthtech startup where I worked, (during the peak of the pandemic) the engineers didn’t need to hear about provincial politics; they needed to know the exact Denominator/Numerator logic for clinical reporting. And because I understood the mess of the clinical data, I could provide the technical specifications that allowed the product/tools to be built without second-guessing.</p><p>The journey from Conception to Completion is rarely a straight line. It’s a messy, iterative process that requires a thick skin and a commitment to evidence.</p><p>If you want to build products that actually change user behavior, whether it’s helping a doctor spend less time on paperwork or helping a newcomer build their first credit score, you have to start with the data. You have to be willing to look at the fragmented logs and the confusing spreadsheets and say, “There is a story in here. Let’s go find it.”</p><p>Because at the end of the day, opinions will vary, strategies will shift, and roadmaps will be debated. But in the high-stakes world of product development, one thing remains constant: Data doesn’t lie.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c5ae7c81bb58" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[From Data to Daily Life: How to Understand Your Diabetes Risk Factors.]]></title>
            <link>https://rachaelo.medium.com/from-data-to-daily-life-how-to-understand-your-diabetes-risk-factors-1ead60380716?source=rss-f917ac799823------2</link>
            <guid isPermaLink="false">https://medium.com/p/1ead60380716</guid>
            <dc:creator><![CDATA[Rachael Olaniyan]]></dc:creator>
            <pubDate>Wed, 26 Feb 2025 17:30:05 GMT</pubDate>
            <atom:updated>2025-02-27T09:28:30.795Z</atom:updated>
            <content:encoded><![CDATA[<p>Diabetes isn’t just a distant threat; it’s woven into the fabric of our daily lives. But how do we unravel those threads? Many of us are drowning in data but starved for understanding. This post is your lifeline, bridging the gap between numbers and real-world implications, and empowering you to take control of your diabetes risk.</p><p>Diabetes isn’t just a clinical term; it’s a daily reality for millions, shaping lives with complications like heart disease, vision loss, and kidney disease. It’s a condition that can silently erode quality of life, often going undiagnosed until it’s too late. To better understand this pervasive health challenge, I turned to data from the Centers for Disease Control and Prevention (CDC). The CDC’s Behavioral Risk Factor Surveillance System (BRFSS) is a crucial resource, gathering insights from hundreds of thousands of Americans annually about their health behaviours. This <a href="https://www.kaggle.com/datasets/alexteboul/diabetes-health-indicators-dataset/data?select=diabetes_012_health_indicators_BRFSS2015.csv#:~:text=code-,diabetes_012_health_indicators_BRFSS2015,-.csv">dataset</a>, derived from the 2015 BRFSS survey, allowed me to explore the intricate relationships between various lifestyle factors and diabetes risk. For a deeper look into the scale of diabetes in America and its impact, I recommend reading this <a href="https://diabetes.org/about-diabetes/statistics">article</a> from American Diabetes Association (ADA). It focuses on the Statistics and key facts about diabetes in the United States.</p><p>The dataset I utilized for this analysis, derived from the 2015 BRFSS survey, presented a unique opportunity to explore these critical health questions. The original, more extensive dataset, which you can explore <a href="https://www.kaggle.com/datasets/alexteboul/diabetes-health-indicators-dataset/data?select=diabetes_012_health_indicators_BRFSS2015.csv">here</a>, provided a wealth of information, albeit in a format that required significant cleaning and transformation. To streamline this process and extract meaningful insights, I employed SQL queries to refine the data, which you can review in my <a href="https://github.com/bukkyy/DIabetes-Indicator-Data-Clean-up-SQL-">GitHub repository</a> . The resulting clean and consolidated dataset, which I’ve made publicly available on <a href="https://drive.google.com/file/d/1qao9ZjJF6uE_WVdcYj6MbOO0OiVEGYrM/view?usp=sharing">Google Drive</a> , allowed me to focus on analyzing the prevalence and impact of various risk factors on diabetes.</p><p>To bring these data insights to life, I visualized the key findings in an interactive <a href="https://public.tableau.com/app/profile/bukky3649/viz/DiabetesHealthIndicators_17405058256100/DiabetesHealthIndicators">Tableau dashboard</a>, which can be explored. This dashboard serves as a dynamic lens through which we can examine the prevalence and patterns of diabetes. Among the 253,680 survey respondents, a striking trend emerged: 52.09% of those with diabetes are female, while 47.91% are male, subtly suggesting a higher susceptibility among women. Further breaking down the data, we observe that 1.83% of respondents are prediabetic, 13.93% are living with diabetes, and a significant 84.24% are non-diabetic. This visual exploration not only underscores the scale of diabetes but also offers a tangible way to understand its distribution across different demographics.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*2Fd593d1PzrRy3dunSAbFg.png" /><figcaption>“Fig. 1: A preview of the multiple dashboards created to explore diabetes risk factors.”</figcaption></figure><h3><strong>The Progression of Risk: Age and Diabetes</strong></h3><p>The age distribution of diabetes paints a compelling narrative of vulnerability that deepens with time. Stepping into the realm of confirmed diabetic patients, we see a striking escalation in prevalence as age advances.</p><p>Among young adults aged 18–24, the presence of diabetes is a mere whisper, affecting just 0.22%. However, this whisper steadily grows into a chorus, with each subsequent age bracket revealing a heightened susceptibility. By the time we reach the 60–64 age group, the prevalence surges to 16.22%, a stark testament to the cumulative impact of time on health. The peak, a crescendo of concern, arrives in the 65–69 age group, where 18.55% of individuals are living with diabetes. This goes on to prove the <a href="https://www.ahajournals.org/doi/10.1161/01.CIR.100.10.1132">article</a> published by the American Heart Association (AHA), which states<em> “</em><strong><em>Diabetes is a particularly strong risk factor for women and the elderly.</em></strong>” Interestingly, the prevalence begins to recede slightly in the later years, suggesting a potential shift in demographics or perhaps the impact of longevity. This age-stratified view offers a poignant reminder of the ever-evolving landscape of health and the critical need for proactive diabetes management as we journey through life.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*oh3ekZ3glFf0kpOIC2YLgg.png" /><figcaption>Fig. 2: Exploring Diabetes and Age: See the distribution of diabetic individuals within various age brackets.</figcaption></figure><h3><strong>Obesity’s Overwhelming Prevalence in Diabetes</strong></h3><p>The scales tip dramatically when we examine the relationship between Body Mass Index (BMI) and diabetes. Within the landscape of diabetic individuals, a staggering 58.16% fall into the obese category, a clear indication of the strong correlation between weight and diabetes. Conversely, those maintaining a normal weight represent a mere 11.11% of diabetic patients, highlighting the protective role of a healthy BMI. The overweight category, while significant, accounts for 30.25% of diabetic individuals, underscoring that even moderate weight gain can increase susceptibility. The underweight category, though small at 0.48%, serves as a reminder that diabetes can affect individuals across the weight spectrum. These figures paint a vivid picture of how weight, particularly obesity, acts as a major player in the diabetes narrative, emphasizing the importance of maintaining a healthy BMI as a cornerstone of diabetes prevention and management. According to the <a href="https://www.niddk.nih.gov/health-information/diabetes/overview/preventing-problems/heart-disease-stroke#:~:text=Obesity%20and%20belly%20fat,chances%20of%20developing%20heart%20disease.">National Institute of Diabetes and Digestive and Kidney Diseases (NIH)</a>, Being overweight or obese can make it harder to manage diabetes and raise your risk for many health problems, including heart disease.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*rUTxGVvRIJYYC6d4QJuFRg.png" /><figcaption>Fig. 3: Obesity Prevalence: Highlighting the prevalence of obesity within the diabetic population</figcaption></figure><h3><strong>Diabetes: A Primary Driver of Heart Disease</strong></h3><p>For those living with diabetes, the data paints a stark picture of compounded risk. It is important to understand that diabetes is a major risk factor for cardiovascular disease. According to a <a href="https://pubmed.ncbi.nlm.nih.gov/31722562/">Pubmed Article</a>, “<strong><em>People with diabetes have a 2–4 times increased risk of CVD compared to people without diabetes</em></strong><em>.”</em></p><p>The high percentage of individuals with a history of heart disease (32.97%) within the diabetic group underscores the powerful connection between uncontrolled blood sugar and cardiovascular health. High blood sugar damages blood vessels and nerves that control the heart, greatly increasing the likely hood of heart disease. Therefore, heart health should be a primary focus for individuals managing diabetes. Compounding this, a significant portion of diabetic individuals also exhibit a history of stroke (31.75%), high blood pressure (24.45%), high cholesterol (22.01%), and obesity or being overweight (17.21%), further escalating their health vulnerabilities. These alarming statistics underscore how uncontrolled diabetes, and daily habits, such as a sedentary lifestyle, diets high in processed foods and saturated fats, chronic stress, and inadequate sleep, can significantly contribute to the development of heart disease. To understand more about these habits and their impact on heart health, I recommend reading <a href="https://www.heart.org/en/healthy-living/healthy-lifestyle">this article</a> from American Heart Association. In stark contrast, the non-diabetic group demonstrates a remarkably low prevalence of these critical risk factors, highlighting the devastating effect of diabetes, and underscoring the importance of preventing and properly managing diabetes.”</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*KpPuAuv8EHZawB_qKoceIQ.png" /><figcaption>Fig. 4: Risk factor disparities: Comparing the prevalence of key factors between diabetes status groups</figcaption></figure><p>In conclusion, the analysis of the 2015 BRFSS data underscores the profound impact of lifestyle and health history on diabetes risk. The data reveals stark disparities in risk factor prevalence between diabetic, prediabetic, and non-diabetic individuals, highlighting the critical role of proactive health management. Based on these findings, I strongly recommend focusing on the following key areas:</p><ol><li><strong>Prioritize cardiovascular health:</strong> According to <a href="https://www.medicalnewstoday.com/articles/diabetes-heart-disease-connection#taking-care-of-the-heart">MedicalNewsToday</a>, Managing blood sugar, blood pressure, and cholesterol levels is the most important way for a person with diabetes to take care of their heart. They can usually achieve this through diet, medication, exercise, and other lifestyle modifications.</li><li><strong>Emphasize weight management:</strong> With obesity and overweight being significant risk factors of diabetes, maintaining a healthy BMI through balanced nutrition and regular exercise is crucial. According to mayo clinic in this <a href="https://www.mayoclinic.org/diseases-conditions/type-2-diabetes/in-depth/diabetes-prevention/art-20047639#:~:text=People%20in%20one%20large%20study,includes%20weightlifting%2C%20yoga%20and%20calisthenics.">article</a>, Engaging in regular physical activity, like brisk walking, swimming, or strength training, helps burn calories and improves insulin sensitivity, further reducing diabetes risk. As <a href="https://www.mayoclinic.org/diseases-conditions/type-2-diabetes/in-depth/diabetes-prevention/art-20047639#:~:text=People%20in%20one%20large%20study,includes%20weightlifting%2C%20yoga%20and%20calisthenics.">stated</a> by a Mayo Clinic staff, people in one large study reduced their risk of developing diabetes by almost 60% after losing approximately 7% of their body weight with changes in exercise and diet.</li><li><strong>Promote healthy lifestyle habits:</strong> Addressing modifiable risk factors like smoking, poor diet, and physical inactivity can significantly reduce diabetes risk. The National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) provides tips for healthy living with diabetes in <a href="https://www.niddk.nih.gov/health-information/diabetes/overview/healthy-living-with-diabetes">this article</a>.</li></ol><p>By embracing these recommendations, individuals can take proactive steps to mitigate their diabetes risk and improve their overall well-being.</p><p><strong>References</strong></p><ol><li>American Diabetes Association. <em>Statistics About Diabetes</em>. Retrieved from <a href="https://diabetes.org/about-diabetes/statistics">https://diabetes.org/about-diabetes/statistics</a></li><li>American Heart Association. <em>Cardiovascular Disease &amp; Diabetes</em>. Retrieved from <a href="https://www.heart.org/en/health-topics/diabetes/why-diabetes-matters/cardiovascular-disease--diabetes">https://www.heart.org/en/health-topics/diabetes/why-diabetes-matters/cardiovascular-disease--diabetes</a></li><li>American Heart Association. <em>Lifestyle Changes to Improve Your Heart Health</em>. Retrieved from <a href="https://www.heart.org/en/healthy-living/healthy-lifestyle/lifestyle-changes-to-improve-your-heart-health">https://www.heart.org/en/healthy-living/healthy-lifestyle/lifestyle-changes-to-improve-your-heart-health</a></li><li>American Heart Association. <em>Why is Cardiovascular Disease the Leading Killer?</em>. Retrieved from <a href="https://www.heart.org/en/health-topics/consumer-healthcare/what-is-cardiovascular-disease/why-is-cardiovascular-disease-the-leading-killer">https://www.heart.org/en/health-topics/consumer-healthcare/what-is-cardiovascular-disease/why-is-cardiovascular-disease-the-leading-killer</a></li><li>Centers for Disease Control and Prevention. <em>Healthy Weight</em>. Retrieved from <a href="https://www.cdc.gov/healthyweight/index.html">https://www.cdc.gov/healthyweight/index.html</a></li><li>Centers for Disease Control and Prevention. <em>National Diabetes Statistics Report</em>. Retrieved from <a href="https://www.cdc.gov/diabetes/data/statistics-report/index.html">https://www.cdc.gov/diabetes/data/statistics-report/index.html</a></li><li>Mayo Clinic. <em>Type 2 diabetes</em>. Retrieved from <a href="https://www.mayoclinic.org/diseases-conditions/type-2-diabetes/symptoms-causes/syc-20351193">https://www.mayoclinic.org/diseases-conditions/type-2-diabetes/symptoms-causes/syc-20351193</a></li><li>National Institute of Diabetes and Digestive and Kidney Diseases. <em>Healthy Eating &amp; Physical Activity</em>. Retrieved from <a href="https://www.niddk.nih.gov/health-information/diabetes/overview/preventing-problems/healthy-eating-physical-activity">https://www.niddk.nih.gov/health-information/diabetes/overview/preventing-problems/healthy-eating-physical-activity</a></li></ol><p><strong>Tech Stacks Used:</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/656/1*xuplMvxEPA_8FlCbRPQ-KA.png" /><figcaption>Fig. 5: The Data Dream Team: A visual lineup of the tech stacks that made the analysis possible.</figcaption></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=1ead60380716" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>