Abba’s girl || Mother of Nations 🌎

Data Scientist || Aspiring ML Engineer || BNSc

[Abba’s girl || Mother of Nations 🌎 @SotonwaSuccess]

Public service announcement 📢

My NHANES ML project is finally complete! 😌
I built a Random Forest model to predict multiple cardiometabolic outcomes:

F1 score: 0.76–0.91
PR AUC: 0.75–0.93

Not perfect, but it works 💅🏽
📂 Full project: [https://t.co/sxEEoNzWpS]

[Abba’s girl || Mother of Nations 🌎 @SotonwaSuccess]

I’m starting a new ML project! 🧵

I’ll be building a system that predicts obesity-related conditions in children and adults using the NHANES dataset.

Conditions like; Obesity, Pre-diabetes, Metabolic syndrome and Cardiovascular disease..and possibly more as the project evolves.

[Abba’s girl || Mother of Nations 🌎 @SotonwaSuccess]

Update: My cardiometabolic risk ML project is now LIVE 🚀

A web app predicting 5 risks: Diabetes, Prediabetes, Dyslipidemia, Metabolic Syndrome & CVD.

Try it 👇
https://t.co/6bDwkV7ArE
(First load ~20–60s)
#HealthTech #AI #DataScience

[Abba’s girl || Mother of Nations 🌎 @SotonwaSuccess]

Trust it's not too late to share my 2025 review.
2025 was a lot!
That's the easiest way to put it, but one thing I'm grateful for is that I ended the year spotting the mindset shifts I needed
Full review: https://t.co/vLXNPTjWCS

[Abba’s girl || Mother of Nations 🌎 @SotonwaSuccess]

@prithewise I have finished tarrying 😂😭
Na now I just wan sleep

[Abba’s girl || Mother of Nations 🌎 @SotonwaSuccess]

@abhitwt C

[Abba’s girl || Mother of Nations 🌎 @SotonwaSuccess]

Me and my Jupyter Notebook😂

[Helen Dee 💡 @_devHelen]

Me: I’ll rest this weekend
Also me: proceeds to open VS Code out of boredom

[Abba’s girl || Mother of Nations 🌎 @SotonwaSuccess]

@graphics_mas Hi
Sure

[Abba’s girl || Mother of Nations 🌎 @SotonwaSuccess]

Public service announcement 📢
I’ve finished the most part of my project 🤭
My model is ready 😌
Using the F1 and PR score, the performance ranges between 0.75 - 0.93
It’s not a perfect score, but it’ll do for now 💅🏽
By the way, the selected model is RF2 - Random Forest

[Abba’s girl || Mother of Nations 🌎 @SotonwaSuccess]

After trying out different models, the best performing model so far is RandomForestClassifier but I’m still not satisfied with its overall performance so I still have to improve the performance.

[Abba’s girl || Mother of Nations 🌎 @SotonwaSuccess]

@Ola_yinka224 @Fimiii06 I told my friend this last week.
The air has gotten dry, but because of how hot the weather is, it’s very subtle
I’ve been drinking water like a fish for over a week now😭😭😭

[Abba’s girl || Mother of Nations 🌎 @SotonwaSuccess]

@tech_girl @in_abuundance I beg to differ though
And it’s not applicable to everyone

[Abba’s girl || Mother of Nations 🌎 @SotonwaSuccess]

I wanna deploy it to an app, but I haven’t learnt model deployment yet, so when I do that, I’ll let y’all know 🤭😌

Oh, I’ll share my GitHub link when I publish it on my GitHub.
Thank you guys 😊

[Abba’s girl || Mother of Nations 🌎 @SotonwaSuccess]

@AdekolaJoy4 @UNICEF Ese ma

[Abba’s girl || Mother of Nations 🌎 @SotonwaSuccess]

Working with @UNICEF is actually my dream job 🙂‍↔️
Those of you who know me, knows how much I love children and how it’s my life’s mission to give them better chances in life 🫠

God willing, I will live my life to make an impact on the lives of millions of children 🤲🏽

[Job Corner @JOBCORNER247]

Don't be shy, just tag the company you wish to work for.

[Abba’s girl || Mother of Nations 🌎 @SotonwaSuccess]

@27thguyyy I use both almost everyday 😂

[Abba’s girl || Mother of Nations 🌎 @SotonwaSuccess]

After trying out different models, the best performing model so far is RandomForestClassifier but I’m still not satisfied with its overall performance so I still have to improve the performance.

[Abba’s girl || Mother of Nations 🌎 @SotonwaSuccess]

Been a minute guys
A lot has been going on.
But the new update is that after creating the model, it had 100% accuracy and I realized it’s because I used the same features in creating the target variables to train the model.
So I had to drop them and retrain the model without them

[Abba’s girl || Mother of Nations 🌎 @SotonwaSuccess]

@JOBCORNER247 @UNICEF

[Abba’s girl || Mother of Nations 🌎 @SotonwaSuccess]

Been a minute guys
A lot has been going on.
But the new update is that after creating the model, it had 100% accuracy and I realized it’s because I used the same features in creating the target variables to train the model.
So I had to drop them and retrain the model without them

[Abba’s girl || Mother of Nations 🌎 @SotonwaSuccess]

Update 🙂‍↔️
I’m done with EDA🥲
Did a few digging around to understand the trend and distribution in the data.
The features are skewed, there are outliers and class imbalance.
But, I’m going to be using tree models for it, and they’re not as sensitive to skew mess & outliers

[Abba’s girl || Mother of Nations 🌎 @SotonwaSuccess]

@aproko_doctor A large capacity humidifier🥲

[Abba’s girl || Mother of Nations 🌎 @SotonwaSuccess]

I’m open to suggestions if you have any 😊

[Abba’s girl || Mother of Nations 🌎 @SotonwaSuccess]

Update 🙂‍↔️
I’m done with EDA🥲
Did a few digging around to understand the trend and distribution in the data.
The features are skewed, there are outliers and class imbalance.
But, I’m going to be using tree models for it, and they’re not as sensitive to skew mess & outliers

[Abba’s girl || Mother of Nations 🌎 @SotonwaSuccess]

Update guys.
I’m done with the data gathering, cleaning and preparation.
I think I ended up downloading about 8 datasets per cycle
Insulin, plasma glucose and HbA1c has also added.

[Abba’s girl || Mother of Nations 🌎 @SotonwaSuccess]

I need the model to be generalized, even in extreme medical conditions so I decided to leave them as they are.
The data is already encoded, so there’s no need for that.
Up next, Model Development 🙂‍↔️