Working on an exciting Engineering project Now I begin to see the correlation between Data Analytics and Mech Engineering Anticipate🔥

See the South Africa we let go to the World Cup ahead of us

If you're looking to start your tech or career journey and you're still in uni in Nigeria. Pleaseeeeeeeee work on your communication skill like learn how to put your words together train yourself 🙏🏾 you could bag a life changing job and still lose it due to communication issues.

Ahhh😂

Python simplified for you 🚀

Happy New Month. Here’s to a month that rewards hard work, opens new doors, and brings more jobs. June, we’re ready. 🙏

Happy New Month! To everyone building in tech, learning new skills, applying for jobs, freelancing, launching startups, contributing to open source, and chasing opportunities, I hope this month brings you closer to your goals. May your applications get responses, your interviews turn into offers, your projects gain visibility, and your hard work open doors you never expected. Wishing every techie, founder, designer, developer, marketer, analyst, and creator a productive and successful month ahead. Happy New Month. Let's build, grow, and win together.

Even Nottingham Forest man😭😭

Admin just had to cook😂😂

😂😂😂😂😂😂😂 Arsenal

Andrej Karpathy’s advice for beginners getting into AI/ ML: “Put in 10,000 hours of work.”

Actually trying to make sense of my lectures from today Heat and Mass transfer

I just wrapped up an intensive project with WorldQuant University, where I built an end-to-end predictive pipeline to forecast air quality (PM2.5 levels) in Nairobi. Unlike standard tabular datasets, time-series data requires an entirely different engineering and modeling mindset. Here is a breakdown of the technical milestones and core takeaways from this project: 1. The Data Pipeline (NoSQL to Pandas) Before modeling, you have to ingest the data. I worked with semi-structured air quality metadata stored in a MongoDB database. Connected via Python drivers to query specific collection metrics. Processed and flattened nested JSON structures into a structured Pandas DataFrame. Handled irregular sensor frequencies by resampling data into a fixed hourly index and addressing missing values to maintain continuity. 2. Feature Engineering Through "Lags" In time-series, you often do not have external features. The past becomes your input. By shifting the timestamp index, I engineered "lag features" (using the prior hour's air quality to predict the next hour). This self-supervision transforms raw sequences into a supervised learning matrix. 3. Deciphering the Math (ACF and PACF) To determine how many historical lags our Autoregressive (AR) model actually needed, I utilized: Autocorrelation Function (ACF): To measure the total correlation between current and past data points. Partial Autocorrelation Function (PACF): To strip away the "noise" of intermediate steps and isolate the direct impact of a specific past hour on the present. 4. Why Traditional Validation Fails (Enter WFV) One of my biggest takeaways was why we cannot use standard random train-test splits or K-Fold Cross-Validation for time-series. Doing so causes "data leakage" because a model cannot look into the future to predict the past. Instead, I implemented Walk-Forward Validation (WFV). This mimics real-world deployment: the model predicts the next time step, tests against the actual result, folds that real result into its training history, and moves forward one step at a time. Building this with ⁠statsmodels⁠, ⁠pymongo⁠, and ⁠scikit-learn⁠ has completely reshaped how I approach chronological data.

Another day to learn, grind and stay focused.

Walahi I hope you win. It's long overdue, to be honest.

Ohh it has been lifted now React if you see my post 🤚🙃

Long story short cut short I am in cohort 9 Data with Danny Thank you to @iam_daniiell

Good afternoon @iam_daniiell First I plead you take out your time to read through My name is Franklin..I am currently a 300 level MechE student I dare not start without acknowledging your voice and work in this data space I am look to join your cohort for this year from June to November but the barrier is the price… I think just maybe I might be able to afford half of it.. I am not new to this space.. as a matter of fact I hold an Oracle Data science cert and a few other Data Analyst certifications.. but it is not enough.. I have a very good knowledge of excel python and SQL to a level…but I believe mentorship from an industry expert like you is what will truly kickstart solid progress… I pray for a positive response And if this gets to your TL Help repost until he sees it

I don’t really understand why.. but okay