Services
Three productized offerings with concrete deliverables, timeline, and what I need to start.
Process
Scope → MVP → hardening/deploy. Keep it simple and ship.
Step 1
Scope
Align on goal, users, data access, constraints, and deadline.
Step 2
MVP
A usable first version you can review early and iterate on.
Step 3
Hardening & deploy
Performance, QA, basic monitoring, docs, and deployment.
Dashboards & data apps (Dash/FastAPI)
Interfaces to explore, decide, and share: dashboards, internal tools, and small data products.
Deliverables
- Dashboard/app with 1–3 key views
- Clean data model + integrations (files/DB/API)
- Reproducible deploy (Docker) + handoff
- Short docs + maintenance checklist
Typical timeline
2–3 weeks for a typical scope (1–3 key views + clear data sources).
1-week MVP if scope is tight and data access is ready.
What I need from you
- Goal + target users (roles)
- Examples of decisions/questions it must support
- Sample data first, then credentials/access
- Constraints (hosting/compliance) + deadline
Pipelines / scraping / automation
Reliable ingestion and transformation: scrapers, scheduled jobs, and clean outputs for analysis or apps.
Deliverables
- Pipeline (ingest → validate → transform) with basic tests
- Scraper(s) with rate limits + change handling
- Clear outputs (tables/CSV/Parquet/API) for your use case
- Logging + basic alerts (failures, freshness)
Typical timeline
2–3 weeks for a typical scope (1–2 sources + stable output).
1-week MVP: one end-to-end flow for one source + one output.
What I need from you
- Source URLs/APIs + permissions
- Output spec (fields, granularity)
- Frequency/SLA (daily/weekly) + consumers
- Destination (DB/object storage) + access
Forecasting / applied econometrics
Forecasting and practical causal work: projections, sensitivity, policy evaluation, uncertainty.
Deliverables
- Baseline + improved model with backtesting
- Short report: assumptions, metrics, uncertainty
- Scenarios/what-if + sensitivity checks
- Notebook + reproducible pipeline (when needed)
Typical timeline
2–3 weeks for a typical scope (clean history + clear horizon).
1-week MVP: baseline forecast + backtest + next steps.
What I need from you
- Historical series (and covariates if available)
- Target definition + horizon
- Known events (promos, policy changes, shocks)
- How the forecast will be used (ops vs strategy)