<?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 Tedd on Medium]]></title>
        <description><![CDATA[Stories by Tedd on Medium]]></description>
        <link>https://medium.com/@kimanited73?source=rss-c39869b82b38------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*5_bpRNlYopPENSKhCufkeQ.jpeg</url>
            <title>Stories by Tedd on Medium</title>
            <link>https://medium.com/@kimanited73?source=rss-c39869b82b38------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sun, 22 Nov 2026 14:51:44 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@kimanited73/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[Open Weight Models]]></title>
            <link>https://medium.com/@kimanited73/open-weight-models-f504be677b1c?source=rss-c39869b82b38------2</link>
            <guid isPermaLink="false">https://medium.com/p/f504be677b1c</guid>
            <category><![CDATA[openai]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[open-weight-model]]></category>
            <dc:creator><![CDATA[Tedd]]></dc:creator>
            <pubDate>Wed, 13 Aug 2025 13:39:44 GMT</pubDate>
            <atom:updated>2025-08-13T13:39:44.500Z</atom:updated>
            <content:encoded><![CDATA[<p><strong><em>What are they, and why should you care?</em></strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/655/1*8_kKauoW0RsAGY0r-uiHLg.png" /><figcaption>An Artificial Neural Network</figcaption></figure><p><strong>Open-weight models</strong> are AI systems whose parameters or weights that define their learned knowledge are made publicly available for <em>personal download</em>. Unlike closed models accessed only via API, open-weight models can be run locally, fine-tuned for specialized tasks, or integrated directly into offline systems. Making them powerful tools for developers, researchers, and organizations seeking independence from third-party providers.</p><p>Just this month(<em>Aug 2025</em>), OpenAI made headlines with the <strong>launch of GPT-OSS, its first open-weight release since GPT-2 in 2019</strong>. The move signals a notable shift in the company’s approach, after years of focusing on closed, API-only access.</p><p><strong>GPT-OSS</strong> represents a turning point as it shows renewed openness to sharing cutting-edge capabilities, providing developers more control over model usage, cost, and customization.</p><h3><strong>From GPT-2 to API-Only Access: How We Got Here</strong></h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1017/1*GLjUBDqYB5f5p0XujyeBRA.png" /></figure><p>When OpenAI released GPT-2 in 2019, the AI world was already buzzing. GPT-2 was revolutionary at the time, but its full release was deliberately delayed due to concerns about misuse.</p><p>A controversial decision that drew both praise for caution and criticism for stifling openness from the global developer community.</p><p>Over the following years, they moved towards an API-first model. This decision was mainly driven by multiple factors:</p><ul><li><strong>Safety and control</strong> &gt;&gt; There is still a lot of misuse of AI tools, this made it easier to monitor and limit misuse.</li><li><strong>Monetization</strong> &gt;&gt; A steady supply of subscriptions, APIs provided predictable, recurring revenue.</li><li><strong>Rapid improvement cycles &gt;&gt; </strong>Iterative development: central control made it easier to update models.</li></ul><h3><strong>Changing AI Landscape: 2019 vs. 2025</strong></h3><p>In 2019, AI was dominated by a handful of research labs and startups. Most state-of-the-art models came from well-funded Western companies.</p><p>Training models required massive compute resources, making it largely inaccessible to individuals or smaller organizations.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*LDIrExHSGYRuqsXgytD_kg.jpeg" /><figcaption>AI Usage 2023</figcaption></figure><p>By 2025, the environment is far more competitive and decentralized:</p><ul><li><strong>Hardware costs</strong> have dropped, and cloud GPU rental is more widely available.</li><li><strong>Open-weight releases</strong> from other players (Meta’s LLaMA, Mistral, Stability AI) have normalized open access to high-quality models.</li><li><strong>Specialization</strong> has increased; many models are now tailored for specific industries or languages.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*08LFSvwCUh-gRT8ZWNIdCw.avif" /><figcaption>Developers at Andela</figcaption></figure><h3><strong>Why Open-Weight Models Matter — Especially for African Developers</strong></h3><p>Open-weight models change the equation for developers outside the traditional AI hubs. This allows:</p><ul><li><strong>Offline operation &gt;&gt; </strong>vital where internet access is unreliable.</li><li><strong>Local language fine-tuning &gt;&gt; </strong>enables AI to serve communities in <strong><em>Kiswahili</em></strong>, <em>Yoruba, &amp; Zulu</em>, without converting to English and back.</li><li><strong>Data sovereignty &gt;&gt; </strong>keeping sensitive datasets within local infrastructure. Allowing places like clinics and hospitals to gain secure access</li><li><strong>Cost control</strong> &gt;&gt; removing recurring API fees, infrastructure, and expertise are the new cost.</li></ul><p><strong>For African developers, this means building AI that reflects local needs, without depending entirely on global providers’ infrastructure or policies.</strong></p><h3><strong>We can build for Africa, in Africa, by Africans</strong></h3><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=f504be677b1c" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[When to use A.I?]]></title>
            <link>https://medium.com/@kimanited73/when-to-use-a-i-e3a6d0ec42eb?source=rss-c39869b82b38------2</link>
            <guid isPermaLink="false">https://medium.com/p/e3a6d0ec42eb</guid>
            <category><![CDATA[data]]></category>
            <category><![CDATA[ai]]></category>
            <dc:creator><![CDATA[Tedd]]></dc:creator>
            <pubDate>Wed, 06 Aug 2025 06:28:24 GMT</pubDate>
            <atom:updated>2025-08-06T09:11:00.076Z</atom:updated>
            <content:encoded><![CDATA[<p>To use or not to use, that is the question.</p><p>Currently A.I is the largest buzzword that i have had the pleasure of interacting with. It appears in high level meetings all the way to discussions with the security guard as you head out of work.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*pFxg_Dg7O_sXUZSpcrBt0Q.png" /></figure><p>Every company and their competitor is rushing to hire ML engineers and AI developers to bring their company right up to the bleeding edge of today’s technology. Investing millions, more than some countries’ GDP, all to be part of the AI revolution. Launching internal/external chat bots, using AI models to predict every possible customer grouping and even introducing AI to reply to emails.</p><p>So many possibilities and varying use cases; its hard to consider a problem, that cant be solved by AI. We rarely pause and ask ourselves is it necessary?</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*nJww6KlU6I6V1xRk6PCFUg.png" /></figure><p>I recently had the opportunity to take several courses on AWS Bedrock which covered this concept. I believe its handy for any developer or anyone with interest in AI to understand the current limitations of what machine learning(at the time of writing) is capable of.</p><p>We wont dive into the history or functionality of AI in this article but we will quickly skim through which scenarios ML can be of use. As a visual learner, I abstracted the take-away of the course in to a simple flow diagram.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/279/1*sVJofrNPDp46A9x6BBi1nA.png" /></figure><p>Now that we have a high level overview of how we should evaluate a problem or use case before utilizing machine learning, let’s take a look at the finer details.</p><h3>The Nitty-Gritty</h3><p>Machine learning is appropriate when the problem is <strong>clear and quantifiable</strong> and model predictions align with business objectives and success criteria. A good starting point is to identify a <strong>problem rich in data</strong> that traditional methods cannot solve.</p><p><strong>Reasons to use ML:</strong></p><ul><li>Requires <strong>complex logic</strong> (e.g., generating personalized recommendations).</li><li>Requires <strong>scalability</strong> (handling millions of requests per second).</li><li>Requires <strong>personalization</strong> (delivering tailored outputs at scale).</li><li>Requires <strong>responsiveness</strong> (providing results within seconds under heavy demand).</li></ul><p><strong>Reasons not to use ML:</strong></p><ul><li>Problem can be solved with <strong>traditional algorithms</strong> (not overly complex).</li><li>Does not require adapting to <strong>new or changing data</strong>.</li><li>Requires <strong>100% accuracy</strong>, which ML typically cannot provide.</li><li>Requires <strong>full interpret ability</strong>, which ML models may lack.</li></ul><p><strong>At the end of the day, AI is not a magic wand but a tool.</strong></p><p>One that works best when applied with <strong>clarity</strong>, <strong>purpose</strong>, and <strong>precision</strong>. The real challenge is not whether AI can be used, but whether it <strong><em>should</em> </strong>be used.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e3a6d0ec42eb" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Closing the Year Recap]]></title>
            <link>https://medium.com/@kimanited73/closing-the-year-recap-533f6e4ac172?source=rss-c39869b82b38------2</link>
            <guid isPermaLink="false">https://medium.com/p/533f6e4ac172</guid>
            <category><![CDATA[usage-analytics]]></category>
            <dc:creator><![CDATA[Tedd]]></dc:creator>
            <pubDate>Mon, 02 Dec 2024 08:36:43 GMT</pubDate>
            <atom:updated>2024-12-02T08:36:43.830Z</atom:updated>
            <content:encoded><![CDATA[<p>Hi Guys, it’s been a while since I’ve done one of these. The previous quarters have been quite hectic and although I kept track of the staff’s BI usage, I did not publish any thoughts about it.</p><p>Now that we approach the end of the year, I think its useful to see where people’s minds are, and exactly what data the staff and managers are interested in.</p><p>So let’s dive right in:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ReipN5lZkRrlduzFUxoQVQ.png" /></figure><p>Our first graph looks into the general level of activity across both organizations. We can see a spike in activity during the first week of November coming off a historical low during the end of October.</p><p>We had the largest amount of exports within that first week, which was sustained over the entire month, seems like there is a spike in report-making activity going on. We can get a better insight into exactly what’s being downloaded later.</p><p>Secondly we have the login activity tracker.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*USQfeFRjKFl4miKP-gAHXw.png" /></figure><p>If you are familiar with the previous one, you might have noticed a change in color.</p><p>I decided to differentiate between the login(Green) and logout(Red) times to see how long staff spend on the site. However, no one ever logs out<em>, me included</em>, and that explains why there is no ‘red’ line.</p><p>However, the login time stays around the same, correlating with the same period most staff arrive to the office.</p><p>Thirdly, our most viewed cards graph.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*l7mo4RIb7WdFuEB9kD3JUg.png" /></figure><p>There’s quite a spread of activity in November, with majority of the focus around cards concerned with overdue or outstanding accounts.</p><p>A focus in debt collection shows that our team is focused on the financial position of the company, trying to sort out its credit and debit balances as we approach the end of the year.</p><p>There is always pressure for registered company’s to present a glowing closing financial position as these tend to be published to the public, and more importantly <strong>investors</strong>.</p><p>Next we have,</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*2SApazzey9miP-g68e1hlg.png" /></figure><p>As expected the Technical Non life team is leading, followed by the fiance team, the life team and finally the IT team. However the is increasing interest by managers as 2 of them have managed to surpass at least 2 departments in their level of activity.</p><p>This is off a training conducted for the Tz subsidiary in best practice usage of DOMO, ideally i will host another refresher training in January to get all the new staff and the current team up to speed with DOMO.</p><p>Something to note is that the departmental activity is aggregated across both organization, i.e technical is both countries Technical team.</p><p><em>We will begin differentiating as of next year</em></p><p>Then we have,</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ac9dochyi63EkOiVtJnUtw.png" /></figure><p>Now this is the download log. As mentioned in our look in general activity, the number of exports and downloads in the previous month could be indicative of reporting season.</p><p>Lets find out exactly what was being downloaded:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*e7EWLJeci4LiPhJR34AVPQ.png" /></figure><p>Quite the range of topics, but as we noticed earlier with the card views. The focus in on reporting and balancing of our books. An example is the ~3rd most downloaded card “CREDIT_DAYS_OTP” is concerned with how long its taken for the company to receive premium from a cedant.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*b703W6lz_TxlowLmijDuSg.png" /></figure><p>Lastly is a quick overview on the most popular dashboard for staff to visit. This seems quite self explanatory going off the other graphs, there is a focus on our financial position, ongoing reports and a keen interest on our receivables.</p><p>I’ll look into a way of segregating the 2 organizations to see how different the activity can be, stay tuned for the last article of 2024 some time in January.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=533f6e4ac172" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Human in the Loop Training]]></title>
            <link>https://medium.com/@kimanited73/human-in-the-loop-training-59e8f53bdc87?source=rss-c39869b82b38------2</link>
            <guid isPermaLink="false">https://medium.com/p/59e8f53bdc87</guid>
            <category><![CDATA[user-analytics]]></category>
            <category><![CDATA[machine-learning]]></category>
            <dc:creator><![CDATA[Tedd]]></dc:creator>
            <pubDate>Tue, 12 Nov 2024 06:21:06 GMT</pubDate>
            <atom:updated>2024-11-12T06:26:12.663Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*lKLX-tuIoyzvytkqOm6A_g.png" /></figure><p>Including human actors into AI training</p><p>This is a new strategy in the machine learning space, aiming to integrate human input and expertise into the training cycle of A.I models.</p><p>The aim is to create a collaborative approach, to achieve what neither a machine or a human can do alone, but what they can do together.</p><h3>So how does this work?</h3><p>Humans or mostly researchers actively provide input to the training, evaluation and operation of a machine learning model. They do this by providing guidance, feedback and annotations.</p><p>This is a prevalent method of training customer facing models as it provides them with invaluable insight into corpus of information.</p><p>Some common means of doing this is through:</p><p>1. Providing labels for training — through data annotation or data labelling side gigs that have become prevalent online</p><p>2. Evaluating performance of ML models — whenever you provide feedback on a response or image from ChatGPT or Gemini</p><p>3. Providing feedback to ML models — this happens in 2 main ways</p><p>a. Active learning — The model can rely on a human actor to label data for improve the efficiency of labeling</p><p>b. Reinforcement learning — Humans provide feedback on the trail and error learning process, allowing the model to learn more effectively</p><h3>Why am I telling you about this?</h3><p>The main impact of this strategy is imprinting on the ML model, the biases and assumptions the AI should <em>consider or ignore</em>, as it improves the adaptability and real-world integration of the model.</p><p>This will become more prevalent as Multi-Modal Models that can process and respond in speech become readily available to the large. In a nutshell, to allow AI to become more human-like.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=59e8f53bdc87" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Reviewing the End of June]]></title>
            <link>https://medium.com/@kimanited73/reviewing-the-end-of-june-0e3613c12000?source=rss-c39869b82b38------2</link>
            <guid isPermaLink="false">https://medium.com/p/0e3613c12000</guid>
            <category><![CDATA[usage-analytics]]></category>
            <dc:creator><![CDATA[Tedd]]></dc:creator>
            <pubDate>Mon, 01 Jul 2024 08:28:23 GMT</pubDate>
            <atom:updated>2024-07-01T08:28:23.402Z</atom:updated>
            <content:encoded><![CDATA[<p>Monthly Usage Patterns at EARe</p><p><strong>Logins and Activity:</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*FrFmuCltV52PC7n4tGXjYA.png" /></figure><ul><li>A trend emerges when reviewing monthly usage: <strong>the beginning and end of months</strong> are the most active periods.</li><li>The <strong>final week</strong>, particularly coinciding with the <strong>quarter’s end</strong>, sees a significant increase in logins, downloaded cards, and even shared cards.</li><li>There’s also greater variation in the most viewed cards each month. For example, in June, users focused on reversed contracts followed by overdue account details.</li></ul><p><strong>Login Time Patterns:</strong></p><ul><li>The monthly login pattern continues the weekly trend, with most logins occurring during lunch hours, specifically between <strong>8 AM and 10 AM</strong>.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*x1S_JKHrEjsDBfY8PFweSg.png" /></figure><p><strong>User Groups:</strong></p><ul><li>Non-Life and Finance departments are the most active user groups, with <strong>more than double the activity</strong> compared to the third-place department. This aligns with these departments having the majority of EARe’s staff.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*56gKVVPm_ZHFy4kpdWB4_A.png" /></figure><p><strong>Dashboards:</strong></p><ul><li>The most active dashboards reflect the previous month’s focus. Finance, Actuarial, and Credit Control dashboards lead in usage.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*JlH_ixPyrm8uF0mi9fjSWA.png" /></figure><p><strong>Downloads:</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Ve3hqAwGxOZbdNEZY_ihPA.png" /></figure><ul><li>Download activity exhibited a clear difference between the end of May and June. The <strong>end of May saw minimal downloads</strong> compared to a significant increase by the end of June.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*zUplA8u8dGelujBLVoOEaA.png" /></figure><p>The busiest periods are at the beginning and end of months, especially when coinciding with quarter ends.</p><p>User groups reflect departmental sizes, with Non-Life and Finance leading in activity. Download and dashboard usage patterns also demonstrate a focus on specific tasks at different points in the month.</p><p>Finally, the login time pattern suggests users primarily access the system during their morning routines.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=0e3613c12000" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Tha end of June in Data]]></title>
            <link>https://medium.com/@kimanited73/tha-end-of-june-in-data-8d782b202759?source=rss-c39869b82b38------2</link>
            <guid isPermaLink="false">https://medium.com/p/8d782b202759</guid>
            <category><![CDATA[usage-analytics]]></category>
            <dc:creator><![CDATA[Tedd]]></dc:creator>
            <pubDate>Fri, 28 Jun 2024 08:47:25 GMT</pubDate>
            <atom:updated>2024-06-28T09:09:38.087Z</atom:updated>
            <content:encoded><![CDATA[<p>We’ve changed the format of these reports. We are keeping them more concise and monthly to ensure readability. There’s no need to review usage weekly while the business cycle is measured in months.</p><p>We can always review a week if there is a need, after this week the format of the reports will change to collect data monthly. Let us dive into our last weekly recap.</p><p>There has been steady interest in existing cards this week, but that has not affected the main working hours(Time when Domo sees the most activity).</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*--gsP38P-vkUI7JSIvMVpg.png" /></figure><p>Login activity remained consistent this week.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*FERhDU8TJPWrO2V-kl62eQ.png" /></figure><p>The Reversal Verification Card was a key focus this week. The team ensured accurate data entry, figures and details, into the system.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*0fi81RBbgVMhv0ZwsKauIw.png" /></figure><p>The Life department has surpassed the Finance team in BI tool usage this week.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*SZzGnQW0nJsxvhuKmN9XRw.png" /></figure><p>An interesting trend emerged this week: a positive correlation between the number of cards created and the number downloaded.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*I348Dfsu9rnJDtO2QhEWHA.png" /></figure><p>Downloads primarily consisted of aging reports, balance reports, and most notably, the Reversal Verification Card.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*qgX0KIv30yi69XUh7bYY9w.png" /></figure><p>The most frequently used cards were Finance, Credit Control, and Actuarial, all aligned with the organization’s financial health.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*dF6ZzjeaIk3-HPTSnc3hgQ.png" /></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=8d782b202759" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Second Week of May]]></title>
            <link>https://medium.com/@kimanited73/the-second-week-of-may-d2086af7ebea?source=rss-c39869b82b38------2</link>
            <guid isPermaLink="false">https://medium.com/p/d2086af7ebea</guid>
            <category><![CDATA[user-analytics]]></category>
            <dc:creator><![CDATA[Tedd]]></dc:creator>
            <pubDate>Tue, 14 May 2024 10:08:53 GMT</pubDate>
            <atom:updated>2024-05-14T10:08:53.424Z</atom:updated>
            <content:encoded><![CDATA[<p>Almost at the halfway point in May</p><p>Hi Guys, welcome to my weekly recap!</p><p>Trying out something new with the start, hope it gave you the impression of a YouTube channel.</p><p>Over the previous week we had a spike in activity, the most interesting part was that the majority of said activity was logged on Saturday. Who was logging, viewing, and creating cards over the weekend, turns out it was the IT department making changes to the cards.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Q6EFOcsC7vWpd_jAnoAd3A.png" /></figure><p>If we look at the week in focus, most of the activity took place on Wednesday, keeping in mind that there was an AGM taking place on Friday which was a national holiday. We can see that on Monday some cards were shared, setting the stage for the rest of the week.</p><p>Nothing to write home about the current state but that could change as we go deeper into the tracking.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*13I5EWfMIHRcwFLaJ4r82w.png" /></figure><p>When we look into the most viewed cards, we can see that the number of specific card views still remains low with the highest being 6 views, focusing on Gross Balances. The other cards with the most views were focused on claims, life business, and expired contracts. The focus is still on financial position.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*g3tiPYy-9S5eYRqFFflAxQ.png" /></figure><p>Would it be advantageous to assist the team to have a more varied view or would having a focused insight into the financial or we want them to track other movements. Our data is primarily focused on similar data but we will see/discover a way forward after some vocal consultation.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*K_-242w_EBOeV3DlVpkqgQ.png" /></figure><p>Moving on the most active user group was as usual the Non-life department, followed by Life then Finance, a triad we’ve began noticing over the previous weeks. We do have another manager showing up on the tracker but with only one view. We can potentially accredit this to a mistaken log in or the staff member just taking a quick look</p><p>Looking at the download activity log, there is an increase in number of created cards but it comes along with a drop in downloaded and exported cards.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*eu9Zq8XohsqLRN-ObemttQ.png" /></figure><p>The most viewed card is the most downloaded card! What a connection! Who would have though, jokes aside its good to see a link between the cards. There was also a number of downloads for the premium register and outstanding cash call refunds, investigating how much our customers owe us.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*LU5gyiEa16uBav6IxZ7edg.png" /></figure><p>The most active dashboard was fiancé followed by credit control then actuarial. There was also some activity in the C-suite dashboards along with risk take a gander on how things are progressing.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*eAF7GVDPeHAp49gJGdkCmg.png" /></figure><p>And for the section we were all waiting for the time login tracker, I’m using this to investigate when most people in the company log in into Domo. Most of the activity takes place in the early hours of the morning. 8 to10 am is when ~70% of the logins occur, a caveat to this: our default browser opens Domo on default as a tab. As the day(s) progress the least activity occurs at 1pm lunchtime and 3pm when staff begin closing up their workstations. This shows an expected frequency of activity, well take a look next week and see if this changes.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=d2086af7ebea" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[First Week of May : What May happen?]]></title>
            <link>https://medium.com/@kimanited73/first-week-of-may-what-may-happen-32773ef2c27e?source=rss-c39869b82b38------2</link>
            <guid isPermaLink="false">https://medium.com/p/32773ef2c27e</guid>
            <category><![CDATA[user-analytics]]></category>
            <dc:creator><![CDATA[Tedd]]></dc:creator>
            <pubDate>Wed, 08 May 2024 14:52:12 GMT</pubDate>
            <atom:updated>2024-05-08T14:52:12.772Z</atom:updated>
            <content:encoded><![CDATA[<h3>First Week of May : What May happen?</h3><p>G’Day Team,</p><p>Apologies for the delayed update this week; I’ve been a bit under the weather, but we’ll persevere. Nonetheless, let’s kick off by celebrating the start of a new month.</p><p>Now, diving into the activity on Domo over the close of April, we observed an interesting trend. Firstly, May 1st was Labor Day, an international holiday, which may have affected our metrics.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*otEMQjYwGrk60r_OU55JaA.png" /></figure><p>It’s apparent that the bulk of activity occurred during the first half of the week, with a significant drop-off during the last two days. Notably, a pattern emerged with most cards being created within the initial two days of the week. This could signify a potential trend worth exploring further.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*0Hl7P0JcCo0rA1tM_fK2pQ.png" /></figure><p>In terms of individual card views, there was a notable decrease overall. The standout cards were “Outstanding MDPS (<strong>Inward Balances</strong>)” and “Overdue Premium Adjustment,” each garnering significant attention. This underscores the importance of addressing overdue accounts and gaining better insight into our outstanding balances.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*u-9wK3c5oJl8zPbqiwVwdA.png" /></figure><p>Regarding Monthly Active Users, there haven’t been substantial changes. However, there’s been renewed interest in accessing Domo, likely for graph creation purposes.</p><p>While the number of exported cards increased, downloads saw a decline throughout the week. Additionally, we introduced four new cards into the system.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*A7foFDCD3_34TnuImQVBgA.png" /></figure><p>The most downloaded cards included “Premium Register Non-Life Inward,” “Sorted Treaty Shares,” and “Gross Balances and Aging Report.”</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*LiorTeO7tNKPltmQBAq-sg.png" /></figure><p>As for the most active dashboards, Finance, Credit Control, and Actuarial stood out. Finance, in particular, received over 100 views, indicating significant engagement, notably from our CFO and CEO.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*sb8lEplQSUwnmA5Djsu6cQ.png" /></figure><p>Let’s keep up the momentum and delve deeper into these insights for continued progress.</p><p>Looking into the next steps:</p><p><strong>User Engagement Strategies:</strong> Develop strategies to increase user engagement with Domo, especially during periods of decreased activity. This could involve targeted communications, training sessions, or highlighting the benefits of using Domo for data analysis and decision-making.</p><p><strong>Content Optimization:</strong> Review the content of the most viewed and downloaded cards to understand what resonates most with users. Consider optimizing existing content or creating similar content to maintain user interest and engagement.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=32773ef2c27e" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[End of April — in the rear view mirror]]></title>
            <link>https://medium.com/@kimanited73/end-of-april-in-the-rear-view-mirror-6d60c2f6bf6d?source=rss-c39869b82b38------2</link>
            <guid isPermaLink="false">https://medium.com/p/6d60c2f6bf6d</guid>
            <category><![CDATA[usage-analytics]]></category>
            <dc:creator><![CDATA[Tedd]]></dc:creator>
            <pubDate>Mon, 29 Apr 2024 06:25:59 GMT</pubDate>
            <atom:updated>2024-04-29T06:25:59.453Z</atom:updated>
            <content:encoded><![CDATA[<h3>End of April — in the rear view mirror</h3><p>The last week of April is synonymous with an increase in pressure as regulatory and board reports are being created.</p><p>This process involves a lot of back and forth with various departments, worse than the other end of quarters as the entire previous financial year is under review.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*xgjCsY5K0yhJACp2w5qHEA.png" /></figure><p>Looking at our first graphs, we see that Monday and Tuesday are high in usage intensity.</p><p>The following days follow a reducing level of interest as most reports have already been downloaded and scrutinized. However, looking at the activity breakdown there’s an overall increase in the number of user log-ins, and will investigate that as we go on.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*DOfuziXB0eNbiB1aCIL_xw.png" /></figure><p>In our card details viewer, we notice that performance based around the underwriting year had the staff all buzzing. Followed by treaty performance over the last 2 years, this demonstrates to us that the main interest was in performance over the past year, which directly correlates to cedant performance which is an important discussion in the AGM.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Uefhw9Sb3N-FRnzhTkU0_w.png" /></figure><p>In our 3rd graph, we see that the department’s use of Domo spiked with the Manager accounts taking less interest in Domo over the week.</p><p>The finance department activity level was slightly beaten by the Life department with 1 extra log-in. The IT department also demonstrated an increase in activity that we rarely get to see.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Ad-gl-Cy7e3IhVAFMzLFMA.png" /></figure><p>Looking at the download activity over the same period, compared with the current month. It shows a slight increase in activity that we hope will be sustained heading into May.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*YvUVKgxFl0jcZsw4DdzVUw.png" /></figure><p>Looking into the most downloaded cards, the first title that steps out of the murk is…..Outstanding CCR! Which is closely followed by the Premium register, Gross Balances, and aging reports.</p><p>All these cards are concerned with the financial position of the company, computation of its debts, and credit status.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*pXCvNQVEfjutr0mFo4gXcQ.png" /></figure><p>To close up our review, we confirmed that the most active dashboards are concerned with financials, credit control, and actual figures. All focused on the company’s fiscal stance. Looking back, this tracker provides us with a retrospective view of the staff and company’s focus over the previous week.</p><h3>Plans for Q2 Data Use Tracking</h3><p>So far, the data suggests several areas for improvement.</p><p>First, I will investigate the unexpected rise in user logins to understand the reason behind it. Then, delve into the decrease in manager Domo usage to identify any usability issues or information gaps.</p><p>Next, use these insights to tailor AGM presentations and reports, focusing on financials, credit control, and CCRs.</p><p>And i will continuously improve the tracker by capturing more data like time spent on dashboards and user search queries. Regularly review this data and adjust Domo dashboards and reports to better serve user needs.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=6d60c2f6bf6d" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Heading towards the end of April]]></title>
            <link>https://medium.com/@kimanited73/heading-towards-the-end-of-april-411094e50b13?source=rss-c39869b82b38------2</link>
            <guid isPermaLink="false">https://medium.com/p/411094e50b13</guid>
            <category><![CDATA[usage-analytics]]></category>
            <dc:creator><![CDATA[Tedd]]></dc:creator>
            <pubDate>Tue, 23 Apr 2024 12:51:53 GMT</pubDate>
            <atom:updated>2024-04-23T13:29:52.524Z</atom:updated>
            <content:encoded><![CDATA[<p>Welcome to the third week of April, the first full week of the month. Let’s see if our overall activity picks up as we get back into full activity and wrap up Q1 reporting. <br> <br> To start we can see a spike in activity with Monday and Tuesday demonstrating the most activity. We also had a spike in created cards and the number of exports.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*VUuBYMAFsckIJNEHO0wJ3g.png" /></figure><p>We could also look into what cards are being generated and viewed within the same period to see how often these reports will be used after they are needed. We have noticed a large spike in the number of cards leading to many becoming bloat-cards(unused but they still take up space).</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*BV3ujLYnWECghH9_PxJaFw.png" /></figure><p>In terms of user activity, we have Finance surpassing the life department with 4 extra views. The activity graph would be key in this section, we would be able to track exactly which cards resulted in this spike of activity.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*dtrcbYctLSPj1dgve9NaMw.png" /></figure><p>The activity graph is pretty interesting, with the fewest downloads happening this week compared to the entire month, hopefully, this is an indicator of increased usage in the cloud platform.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*k69AytD-0PDIIrR7UMRWzA.png" /></figure><p>The most downloaded card this week was the non-life gross balances, with the non-life dumps and premium registers coming a close second.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*CHU2R-NzeZ83uetleFXwGg.png" /></figure><p>Overall the most active dashboard was the finance one with over 100 views, followed by credit control the actuarial</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*pXCvNQVEfjutr0mFo4gXcQ.png" /></figure><p>In summary, the third week of April witnessed increased activity as organizations finalized Q1 reporting. Concerns were raised about unused “bloat-cards,” while Finance emerged as the most engaged department. Despite declining downloads, there’s potential for increased cloud platform usage. Key cards like “non-life gross balances” remained popular. Overall, optimizing card usage and focusing on relevant data is crucial for efficiency in reporting.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=411094e50b13" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>