Bold, fresh, unorthodox ideas on how to land your first data analyst job faster and without wasted effort | Check Featured section
Two actions you can take right now to make you more successful in the data field (for rookies & veterans): 1️⃣ - Ditch posting about boring stuff - No posting about guided projects we've all seen a thousand times already. - No uninspired blocks of code, no complex tool stack diagrams, nothing that is mainly addressed to our fellow data people instead of your target audience. Posts like these put recruiters, hiring managers and prospective clients to sleep. What could you post about instead to stand out? My friend Brian Julius has got you covered! Go through his LinkedIn posts (and find his GitHub account), pick one of the excellent ideas he presents for implementation and run with it. -> There's your shortcut to standing out with building original projects. 2️⃣ - Put more energy into building genuine relationships by adding more value to others - No posting generic content into the void. - No low-effort comments and clout chasing. What could you do instead? Start by reposting Brian's post below so people can - discover his work - contribute relevant DAX datasets Brian is looking for so we can all benefit There you have it. Two easy actions you can take right now to improve your data karma. That's all I have for you today. Now go and conquer the data world, one original post at a time. Your future self (and your network) will thank you.
6x Linkedin Top Voice | Lifelong Data Geek | IBCS Certified Data Analyst | Power BI Expert | DAX Heretic | Data Mad Scientist, mixing BI, R, M, AI, PKM, GIS and DS
A big focus of my July content is testing AI's coding ability in languages used by Data Analysts. I've got solid testing sets in every language but DAX, and could really use the help of the DAX expert community here... There is a great deal discussion and speculation regarding the impact that AI will have on the work done by Data Analysts. Unfortunately, however, there is also a paucity of credible testing rigorously assessing how good current frontier AI models are at coding in the languages most frequently used by Data Analysts. In my testing, I plan to assess: 🔸 ability to accurately code problem solutions across a range of difficulty levels for Python, R, M, and DAX (I don't use SQL much, so will leave that one to those w/ more expertise) 🔸 comparative capabilities of different models, as well as the impact of using custom GPTs vs base models 🔸 effects of different prompting strategies on solution accuracy and efficiency A huge part of executing this type of assessment well is choosing an appropriate set of problems on which to evaluate the models' capabilities. I believe I have very solid testing sets for M, Python, and R, but I'm having difficulty developing a strong set for DAX. A suitable DAX problem set should reflect the following criteria: 🔸 Representative of the types of problems analysts would try to appropriately solve using DAX, but are unique enough that the models likely would not have been trained on that same or a very similar problem 🔸 Developed after April 2024 (the training cut-off for Claude Sonnet 3.5 and Opus 3.5) 🔸 Reflect a clearly discernible difficulty gradient from intermediate to expert 🔸 Ideally have elements upon which solution quality (beyond just accuracy) could be objectively assessed. Examples of such factors include speed of execution, brute force versus more sophisticated algorithms, etc. 🔸 Availability of corresponding data/data model (this one is not a must, as these can be synthetically generated to match the problem, but nice if they already exist...) If you have specific problems you think would meet these criteria, or sources where they could be found, please let me know. As I did for the M code assessment I posted earlier this week, I will make all the problems and data available to the community via my github repo for replication and or additional analysis. I think these are really important questions that warrant careful thought and analysis. Thanks in advance for your assistance and feedback. - every time I've reached out in this way prior, I've been knocked out by the expertise and generosity of this community. #dax #ai #dataanalysis #assessment #gpt4O #openai #anthropic ##claude #powerbi
📌 'nothing that is mainly addressed to our fellow data people instead of your target audience.' If your target audience are our fellow data people, then you want to stay in our little echo-chamber, of course.
Brian Julius
2w
Raphael - As always, thank you so much for your friendship, support, and encouragement - both on this particular project, and in general. To those thinking about getting into content creation, I would offer the following encouragement: 🔸 You have something unique and valuable to say, regardless of your level of experience and expertise. Some of my favorite feeds are from those early in their data careers, honestly discussing their experience in finding their way in this fascinating field 🔸 I have never met a content creator (even those who chose not to stick with it in the long run) who regretted giving it a try. However, I know a LOT of creators, myself definitely included, who regret not starting much earlier 🔸 I honestly believe the most interesting day in the history of human civilization to be a data analyst is today. And tomorrow will be more interesting than today. And Thursday will be... 🔸 Many of the most knowledgeable people in the world - those who write the programs we all use, the books we all learn from, etc. are right here - accessible, and eager to discuss If you're thinking "what will I write about?", as Raphael advises - just start substantively engaging and soon you'll be awash in ideas.
Bold, fresh, unorthodox ideas on how to land your first data analyst job faster and without wasted effort | Check Featured section
2w📌 📌 Here is a great strategy to come up with data projects that are not boring: https://www.linkedin.com/posts/raphaelschagerl_tired-of-boring-portfolio-projects-recruiters-activity-7103013842917408768-O4Q4