Time series data science problems with InfluxDB
summary
Shows how InfluxDB uses Telegraf exec processor to run Python at the edge and produce micro‑batch forecasts for a home‑brew temperature control use case, using double exponential smoothing to get accurate univariate forecasts from very sparse, noisy OCR readings. Covers Flux for cleaning, joins and alerts, Jupyter and Zeppelin connectors that return pandas/ Spark dataframes, customer deployments (85,000+ solar rooftops, Isako’s time‑series lake), and a new Rust/Arrow/Parquet storage engine adding SQL and Parquet interoperability.
AI-written summary · about
chapters (11)
- 0:00Introduction and agenda
- 2:02What is time series data
- 3:45Why time series databases
- 4:48InfluxDB platform overview
- 7:14Flux language capabilities
- 10:24Telegraf for edge forecasting
- 15:57Notebook integrations
- 18:06Python time series libraries
- 21:41Customer use cases
- 24:32New InfluxDB storage engine
- 26:40Resources and conclusion
related talks
30mMethods & Tools For Time Series Data Science Problems With InfluxDB
44mMethods and Tools for Time Series Data Science Problems
23mTime series database: Should I use one?
12mTime to awesome: InfluxDB and the AWS environment
47mBack to the Future of Time Series Data
22mExploring the InfluxDB Ecosystem Integrations, Libraries, and Plugins
12mInfluxDB for Industrial IoT
47mBack to the Future of Time Series Data