Machine learning
Predict categories and numbers from your historical data, with models matched to the problem and tested honestly.
- Classification
- Regression
- Clustering
- Feature engineering
- Model tuning
Services built around your data
From a first exploratory analysis to a deployed model and the website that presents it. Pick one service or combine them.
Time series & forecasting
Forecast demand, sales, energy use or sensor readings with seasonality, trend and uncertainty ranges.
- ARIMA / SARIMA
- Prophet
- Exponential smoothing
- Gradient-boosted lag models
Anomaly detection
Spot faults, fraud-like patterns and unexpected behaviour in streams or historical logs.
- Isolation forest
- Statistical thresholds
- Residual-based detection
- Alert rules
Regression & prediction
Estimate prices, yields, durations or costs, and learn which factors really drive the result.
- Linear & ridge
- Random forest
- Gradient boosting
- Error analysis
Matplotlib simulations
Model physical and statistical systems, then visualise them in clean figures and animations ready for papers, reports and teaching.
- Physics & engineering
- Monte Carlo
- Differential equations
- Animations
- Publication-quality figures
Segmentation & clustering
Group customers, products or readings into meaningful segments without labelled data.
- K-means
- DBSCAN
- PCA
- Segment profiling
Data science & analytics
Cleaning, exploratory analysis, statistical testing and reporting that turn spreadsheets and databases into decisions.
- Data cleaning
- EDA
- Hypothesis testing
- Reports & dashboards
Web design & development
Fast, accessible, mobile-friendly websites and data dashboards with a clean, professional look.
- Business websites
- Dashboards
- Landing pages
- Hosting setup
Tools we work with
A mature, open-source Python stack plus modern web tooling.
- Python
- scikit-learn
- pandas
- NumPy
- SciPy
- statsmodels
- XGBoost
- Matplotlib
- Jupyter
- SQL
- HTML / CSS / JavaScript
- Next.js
- PHP
Have data, or a problem worth modelling?
Tell us what you're working on. We reply within one business day with next steps and a clear scope.