Sfoglia categorie
Esplora
Fiverr Pro
Italiano
$
USD
I will predict customer churn using machine learning in python
Financial Modeling, DCF, Power BI, Churn Prediction
Hi, I'm Amna... I actually enjoy building financial models, which probably makes me a little unusual.
Finance MBA with a Biomedical Engineering background. I build financial models, DCF valuations, Po...
Informazioni su questo servizio
Losing subscribers quietly drains revenue and most businesses only find out after the customer is already gone.
I build machine learning models that flag which customers are about to churn before it happens, so you can step in with retention offers while there's still time.
On a real-world telecom dataset (7,000+ customers), my model correctly flagged 82% of the revenue at risk worth over $1.7M annually if scaled to a 10,000-customer subscription business. The model also surfaces why customers are leaving (contract type, tenure, pricing), so you're not just getting a prediction, you're getting a reason.
What you get:
- A working churn prediction model trained on your customer data (or a sample dataset)
- Clear metrics: accuracy, precision, recall explained in plain business terms, not just jargon
- A ranked list of your highest-risk customers
- A short write-up: what drives churn in your business and what to act on first
- Revenue-at-risk estimate in dollar terms, not just probabilities
Whether you run a SaaS product, a subscription box, a telecom service, or any recurring-revenue business if you're losing customers and don't know why, this tells you.
Linguaggio di programmazione:
Python
Framework:
Scikit-learn
•
Panda
API:
Google Cloud Vision API
Strumenti:
Quaderno jupyter
•
Colab
Il mio portfolio
FAQ
Do I need to provide my own customer data?
Not necessarily — you can share your own customer dataset (CSV/Excel) for a fully custom model, or I can use a representative sample dataset to demonstrate the approach if you're not ready to share real data yet. Either way, the deliverable includes the same model, risk scores, and report.
What format will I receive the results in?
You'll get a ranked list of customers by churn risk (Excel/CSV), a written report explaining the key drivers of churn in plain business terms, and — depending on your package — a revenue-at-risk estimate in dollar terms.
How accurate is the model?
Accuracy depends on your data, but on a benchmark dataset my model flagged 82% of at-risk revenue. I'll share precision, recall, and accuracy for your data — the metric that matters most for churn is revenue caught, not just overall accuracy.
Can you deploy the model or connect it to my live system?
Not currently as part of this gig — this delivers a trained model and analysis on the data you provide, not live deployment or API integration. Happy to discuss that as a custom follow-on if needed.
What data do you need at minimum?
Customer-level data with a churn/not-churn outcome and relevant features — things like tenure, contract type, billing amount, usage patterns, support interactions. The more relevant history you have, the stronger the model.
How long does it take?
Delivery time is listed per package (3/5/7 days), but message me first with your dataset size and format so I can confirm timing before you order — very large or messy datasets may need a custom quote.
Do you offer revisions?
Yes, the number included depends on your package — see the package details above. Additional revisions can be added as a gig extra.
