Alex Lomizov

💰 "Quantifying the value of data and analytics initiatives" Last week, just before Big Data LDN, I attended the 7th edition of the London Data Product Management Meetup. Loved the format with around 30+ people, lots of peer chat to warm up around drinks and small breakout discussions on specific topics. We discussed "Quantifying the £££ value of data and analytics initiatives", and here are 7 key highlights / takeaways: 🎯 Reverse the charge on business users: The data & AI teams should not be responsible for value assessment alone, and they should rely on the requesting business stakeholders to assess expected value and then report on realized one 🎯 Ask top-down questions to assess the negative impact of data:  On top of assessing the positive gain from products, you can also assess how the lack of data, or bad data quality would negatively impact the ability to drive business decisions. A great to identify the critical capabilities where monitoring & robustness are required 🎯 Ask bottom-up questions to look at potential usage:  Don't forget to look at your existing Data, Analytics & AI products to identify other potential use, as that might be a low hanging fruit to deliver more value at marginal cost 🎯 Beware of the potential disconnect between data quality improvement and outcomes:  Improving data quality does not always have a direct impact on better outcomes. So before investing, ensure you have a good understanding of the requirements for the different use cases, and only invest where it really matters. 🎯 Manage stakeholders with prioritized list:  Stakeholders often come with urgent demands, expecting to reshuffle the backlog and as such putting the Data, Analytics and AI teams at risk. By consolidating and maintaining a prioritized list of the different demands (along with expected value & estimated cost), and making it visible, it become easier to make those stakeholders prioritize their new demands versus existing ones (or request additional budget for the team to increase bandwidth). 🎯 Think about business vs technical readiness:  When evaluating the complexity of new use cases and products, evaluate both the technical readiness (a.k.a. the ability to build and deliver) and the business readiness, in other words the potential impacts in terms of training, process change, culture change on the business to actually adopt the product. 🎯 Deliver iteratively / adjust effort vs value / find ways to test and derisk:  It should probably be a given, but adopting agile practices to Data, Analytics & AI products is key, to deliver value iteratively, therefor maximizing chances of adoption, maximizing opportunities to gather feedback and continuously improve, and controlling the potential risks in terms of feasibility, costs, etc. Those are very practical tips that everyone can apply in its own context. What do you think about those advice? Anything to add or change here? #DataProductManagement #DPM #ValueDriven #DataProduct

32

7 Comments

Unleash the Power of Data at T&Pm! 💥 We're on the hunt for a UK Data Science & Measurement Lead to work with Sarah Welch Aggarwal and Tom White to champion our analytics! Got a knack for econometrics & MMM? Love turning data into actionable insights? Thrive in a collaborative & dynamic environment? Join our diverse team at T&Pm and make a real impact - link in comments 🚀 #DataScience #Analytics #Leadership #TandPm

25

1 Comment

We had a great evening at Brighton Data Forum this week. An excellent lineup of speakers exploring the topic of data engineering, focusing specifically on the Databricks platform and Spark framework for big data pipelines and processing. Did you come along? If so, what was your top takeaway from the night? Comment below! If you missed this week’s Forum, you can also watch the full recording here: https://lnkd.in/eUvddnbB #Tech #TechEvent #DataEvent #DataForum #Spark #Databricks

12

1 Comment

It's quite often a challenge to maintain governance over dynamic data schemas. Schema drift, characterised by changes in the structure, distribution, or quality of data over time, can have far-reaching implications on downstream systems and processes. By using Apache NiFi as our data movement and orchestration tool, with a touch of custom code, we’ve created a comprehensive schema drift detection and monitoring solution, allowing us to mitigate the potential impacts of schema changes and enabling prompt remediation when necessary. Read all about it in this week's blog post! https://hubs.la/Q02RbLHx0 #data #analytics #apache #nifi #cloudera #schema #dataquality #datagovernance

11

3 Comments

Really chuffed to have been accepted again thanks DSF. Looking forward to this one. A new topic to talk about with a familiar twist on data visualisation. Best get building those slides. Hopefully see some familiar faces on the day, and if you're interested in coming along, check out the free ballot link below. Enter the ticket ballot below 🎟️ https://ow.ly/FvB550QF1Ot #DSFmayday #datasciencefest #datasciencefestival

27

3 Comments

For my UK based Tableau network. Particularly the (ex) contractors among you. In the years pre-Covid, recruiters contacted me more or less every week with a new Tableau opportunity. Since then, recruiters contact me 1 or 2 times per year. The opportunities I see on this platform, perm or contract, are few and all with low salaries/rates - maybe back how they were in the mid-2000s (not for Tableau, but similar level roles). Now, I'm not looking for a role (just to clear that up!), more interested in the state of the market. What's happening? 1. Is Tableau losing a lot of market share, so far less demand for expertise? 2. Is the work being viewed as low value commodity work, and being pushed offshore? 3. Is everyone now perm and not changing roles? 4. Is all of the former contract work going to consulting firms - i.e. IR35 killed the contract market 100% for Tableau? 5. Has the UK imported so many Tableau people in the last few years, there's excess supply of labour? Something else I haven't considered? Are others out there experiencing the same? I know Hari Sekhon writes a lot about similar issues but in the IT space, so it's not a Tableau only thing. #tableau #consulting #contracting

35

44 Comments

Dear Rising Data Star, I trust you are well. One of the greatest mistakes I made about data, as a rising star, was assuming data quality. I assumed the data provided was correct, ran analysis, generated reports, only to get negative feedback. Then I learnt to question the data and the supplier of data. I learnt to run simple sql statements like Select count(distinct x) unique_x_count, count (×) x_count From table_y Where X represents a field confirmed to have unique attributes. If unique_x_count = x_count then field is unique If unique_x_count < x_count then field is not unique Select distinct lenght(x) From table_y Where X is a field of fixed length If result is a single value then all fields have same length, else lengths are inconsistent. These simple scripts were a life saver in my telecom life. Remember always question your data to confirm quality before you start work. PS: It could get more complex than this sometimes. Don't forget to test quality of reports too before delivery. Omolola The Data Demystifier... May 2024 📸 scenes from my new life...

8

3 Comments

Dear Rising Data Star, I trust you are well. One of the greatest mistakes I made about data, as a rising star, was assuming data quality. I assumed the data provided was correct, ran analysis, generated reports, only to get negative feedback. Then I learnt to question the data and the supplier of data. I learnt to run simple sql statements like Select count(distinct x) unique_x_count, count (×) x_count From table_y Where X represents a field confirmed to have unique attributes. If unique_x_count = x_count then field is unique If unique_x_count < x_count then field is not unique Select distinct lenght(x) From table_y Where X is a field of fixed length If result is a single value then all fields have same length, else lengths are inconsistent. These simple scripts were a life saver in my telecom life. Remember always question your data to confirm quality before you start work. PS: It could get more complex than this sometimes. Don't forget to test quality of reports too before delivery. Omolola The Data Demystifier... May 2024 📸 scenes from my new life...

1

If you haven't seen it yet, why not?? 😁 Check out my first ever attempt at a podcast, introducing The Data Den. Episode One features Sue Geuens, CDMP, MDQM, Director of Data Governance & Product Data at Elsevier. Sue talks passionately about storytelling for the data professional. I'd love to hear everyone's feedback on this podcast so it can get better and better with each episode! This episode features discussions around: 📢 Attributes of a good storyteller 📢 Avoiding buzzwords and bad storytelling 📢 Common mistakes in data storytelling 📢 Crafting a story with the goal in mind 📢 How data storytelling is visual 📢 Potential and finding a mentor 📢 Storytelling tips from a professional 📢 Tech tools for data storytelling 📢 Transforming raw data into compelling stories 📢 Using technical terms with stakeholders 📢 Why storytelling is so important https://lnkd.in/eybFpG4R #datadenpodcast #datastorytelling #data

6

1 Comment