Idősor elemzés Jobs

I need a hands-on data scientist to turn a set of time-stamped records into a working regression model that can reliably forecast future values. The project’s primary objective is to develop a predictive model—specifically, a regression solution built on time series data. Here is what I will provide: • A cleaned, well-structured dataset (CSV) that includes the historical target variable and supporting features. • A short brief on the business context and the key performance metric I care about. What I expect from you: • Exploratory analysis that highlights trends, seasonality, outliers, and any useful lags or engineered features. • A robust regression model suited to time-series behaviour (ARIMA, Prophet, XGBoost, LSTM, or similar—feel free t...

We are looking for a Python developer with experience to build a robust, local pipeline that processes Binance Futures historical data into an ML-ready dataset. The goal is to ingest public data from Binance Vision (aggTrades, all klines, and bookDepth) and output clean, normalized, lookahead-bias-free features stored in Parquet format or DuckDB. Scope of Work & Deliverables 1. Ingestion & Database Setup (Core Foundation) Data Source: Programmatic downloading of historical daily/monthly ZIP files from public (specifically aggTrades, all klines [1m], and bookDepth for BTCUSDT, ETHUSDT, SOLUSDT, XRPUSDT, BNBUSDT). Storage Architecture: Set up a local storage solution using DuckDB or Parquet to handle millions of rows without memory issues. Alignment: Parse and align different ...

Need an experienced Python Quant Developer to build a backtesting and research engine for a crypto futures trading strategy (BTC and ETH perpetual futures). This is a Phase 1 research project only. I am NOT looking for a live trading bot at this stage. The objective is to validate the strategy using historical data, realistic execution assumptions, fees, slippage and performance analysis before considering live deployment.

I’m ready to send you a raw Parquet-formatted financial time-series dataset and I need it returned fully cleaned and model-ready. The work centres on four specific steps: • Handling missing values • Scaling and normalisation • Feature engineering • Producing a consolidated master database for model training Please retain the native frequency and time index, document every transformation, and return both the processed file (CSV or Parquet—whichever keeps the structure intact) and the reproducible code or notebook you used. If you’ve tackled stock or crypto data before, this should be straightforward and quick. Once I verify the dataset loads smoothly into our existing notebooks, the job is done.

I have a daily stock-price dataset that already contains the open, close, high, and low values for each trading session. I need a fully reproducible AutoGluon time-series model that can forecasts price movement.