Browse categories
Explore
Fiverr Pro
English
$
USD
I will do data analysis, machine learning and deep learning
About this Gig
Need to understand your data or build a prediction model?
I'm Shazim. I use Python for data analysis, machine learning and deep learning, delivering runnable Jupyter Notebooks with clear explanations.
I help with classification and regression, including customer churn and numerical predictions.
Packages:
- Basic: Clean and explore 1 CSV, train 1 baseline ML model and evaluate results. Up to 5,000 rows and 20 input columns.
- Standard: Compare 3 ML models and tune the best with up to 10 trials. Up to 20,000 rows and 40 input columns.
- Premium: Standard plus 1 small neural network on the same tabular dataset, training plots and model comparison.
Every package includes:
- Commented notebook and preprocessing steps
- Evaluation metrics and charts
- Saved model, dependencies and prediction example
One dataset and one prediction target per order. Please provide existing target labels.
Tools: Pandas, Scikit-learn, TensorFlow/Keras or PyTorch.
Clustering, forecasting, image classification and deployment require custom scope. Performance depends on your data.
Message me with a dataset sample and your goal before ordering.
Programming language:
Python
Frameworks:
Scikit-learn
•
DeepPy
•
Keras
•
PyTorch
•
Panda
Tools:
Jupyter Notebook
•
OpenCV
•
TensorFlow
•
Colab
My Portfolio
Other Data Science & ML Services I Offer
FAQ
What should I send before ordering?
Send a sample CSV, column descriptions, the target you want to predict and your intended use. Include any required metrics, deadline and preferred framework. I'll review data suitability and confirm the scope before you order.
What dataset sizes do your packages cover?
Basic covers 1 CSV with up to 5,000 rows and 20 input columns. Standard and Premium cover up to 20,000 rows and 40 input columns. Each order covers 1 prediction target. Larger, highly complex or multiple datasets need a custom quote
What files will I receive?
A commented Jupyter Notebook, preprocessing code, evaluation metrics and charts, a saved model, dependency list and example prediction code. Setup instructions explain how to run the agreed workflow. Standard and Premium also include model comparisons.
Do I need to provide a clean and labeled dataset?
You provide the dataset and existing target labels. Packages include handling missing values, duplicates, categorical encoding and scaling where appropriate. Data collection, manual labeling, joining multiple sources and extensive data repair require separate scope.
What do model tuning and the Premium neural network include?
Standard and Premium include up to 10 hyperparameter trials for the best classical ML model. Premium adds 1 small tabular neural network, with up to 3 training runs and 50 epochs per run. This covers the same dataset and target. Pretrained-model fine-tuning is separate.
Can you guarantee a specific accuracy or better deep learning results?
No fixed accuracy is guaranteed. Results depend on data quality, sample size and the prediction task. I evaluate models on held-out data using suitable metrics and explain limitations. The Premium neural network may perform better or worse than the classical models.
Do you also offer clustering, forecasting and image classification?
Yes. These projects are quoted separately after reviewing your dataset and requirements. Share a sample, your objective and available computing resources. Image labeling, large training jobs and paid cloud/GPU costs are not included in the listed packages.
Are a website, API or cloud deployment included?
The listed packages deliver notebook-based model development and evaluation. Websites, dashboards, API integration, cloud deployment and ongoing performance monitoring require a separate custom offer. Any paid hosting or compute costs are agreed in advance.
What counts as a revision?
One revision is one round of changes within the agreed scope, such as chart presentation or output formatting. Basic includes 1; Standard and Premium include 2. An extra round is $20 with 2 additional days. Bugs or missed agreed requirements are corrected as part of the original delivery.
What does the Additional ML Model extra include?
For $35 and 2 additional days, I train and evaluate 1 extra classical ML model on the same prepared dataset and target, then update the notebook and results. This extra excludes hyperparameter tuning, neural networks, new datasets and new prediction targets.
