Door categorieën bladeren
Ontdekken
Fiverr Pro
Nederlands
$
USD
I will do data cleaning, and feature engineering to boost your model
I’m Abdelwahed (call me Abdel), a top-ranked data scientist with 7+ years of experience. I hold two Master’s in Big Data Analytics and Executive Data Science and was recognized as Africa’s 2nd best an...
Over deze dienst
Struggling with messy tabular data or an underperforming model?
I specialize in data cleaning, data preprocessing, and feature engineering for tabular machine learning (classification & regression) using Python, pandas, and scikit-learn.
As a Kaggle Master and data science instructor, I deliver:
- Leakage-safe, reproducible pipelines
- Measurable improvements on Accuracy, F1, AUC, or RMSE
What I Do
- Data Cleaning: missing values, outliers, duplicates, type fixes, encoding & scaling
- Feature Engineering: domain, interaction, and time-aware features (no leakage)
- Reproducibility: pipelines with seeds + clear documentation
Deliverables
- Jupyter Notebook
- Feature dictionary
- Before/after metric comparisons
Who I Help
- Business teams needing analysis-ready data
- ML practitioners & Kagglers improving models
- Academic researchers require transparent results
Send me your dataset size, target column, problem type, and metric, and I'll recommend the best approach or craft a custom feature engineering offer tailored to your needs.
Programmeertaal:
Python
•
R
•
MATLAB
Frameworks:
Scikit-learn
•
SimpleCV
•
keras
•
PyTorch
•
Panda
Tools:
Jupyter-notitieboek
•
opencv
•
tensorflow
•
Excel
•
Colab
•
RStudio
Veelgestelde vragen
Will the features you create improve my model's performance?
Yes. I focus on creating features that are statistically significant and relevant to your target variable.
Will you test the features on a model to check if they work?
Yes. I will evaluate the engineered features using a basic model to ensure they contribute positively to performance.
Will you provide the code for the feature engineering?
Yes. All packages include the code for the engineered features. ONLY the **Premium package** gets a Python script to generate a more useful features for future use.
Do I need to send you my model or just the dataset?
You can send just your dataset. I will apply a basic model to assess the impact of the engineered features. However, if you have an existing model, sharing it will allow for more tailored feature engineering.
Can I request specific types of features to be created?
Absolutely. You're welcome to suggest specific features. While I will incorporate them if feasible, I cannot guarantee their impact on your model's performance.
Can you fine-tune my model after feature engineering?
Model fine-tuning is not included in this gig. However, it can be added as an extra service. Please message me to discuss a custom offer tailored to your needs.
Will you explain how to use these features in my model?
Absolutely. You'll receive a Jupyter Notebook showing how each feature was built and how to integrate them into your ML pipeline.
How do you know which features to create?
I analyze your data and your objective, then design features that are most likely to improve prediction accuracy, including transformations, ratios, and interaction terms when needed.
Can you make a custom offer?
Absolutely. Message me with dataset size (rows × columns), task (classification/regression), preferred metric (e.g., F1, RMSE), and timeline. I’ll recommend the best package or send a custom offer with a scoped plan and price.
