Time Series Analytics with Jupyter
Budget: $10 – $100 USD
I am looking to create some analytic Jupyter reports of time series data. It can be stock market type data, curencies or fundamental/ecomomic data. All we can assume is that is is a numeric timeseries.
I have an API that can return 1 (2 or more) time series in JSON format. The data will be returned as JSON and failure (non 200 range result) must ne handled) that I would like to use to to do detailed analysis and comparison in Jupyter.
The data may or may not be stationary. It must be tested where possible (examples in links below).
The API returns any metadata that is needed - if any.
We need to let the user enter/define the dataset names, frequency, fill type, column (OHLC) and date range. We can add defaults. This is to allow different datasets ands parameters to be easily changed. These can be handled as API request parameters or in Python/R.
All reports must be well commented as in the report example attached.
I would like 3 simple reports to start:
a). Time Series Analysis on one time series displaying month averages with seasonality and moving averages etc.
b). As above but with analysis of relationship, movement, difference/correlation of THE 2 time series
c). Seasonality on one time series
Based on: https://machinelearningmastery.com/time-series-seasonality-with-python/
I would like you to explain about the reports you will create and what features each will have. My list above is only a suggestion. If you use another developers code we need to credit them in the notebooks. It also must be legal to use the code freely - even commercially.
You can use Python or R with Jupyter.
Docker experience very useful. You will assist in getting the notebooks installed and running on our server after they are created. We have Jupyterlabs installed with Python and R.
Thanks
Links:
https://machinelearningmastery.com/time-series-seasonality-with-python/
https://goodboychan.github.io/python/datacamp/time_series_analysis/machine_learning/2020/06/18/01-Predicting-Time-Series-Data.html
I have an API that can return 1 (2 or more) time series in JSON format. The data will be returned as JSON and failure (non 200 range result) must ne handled) that I would like to use to to do detailed analysis and comparison in Jupyter.
The data may or may not be stationary. It must be tested where possible (examples in links below).
The API returns any metadata that is needed - if any.
We need to let the user enter/define the dataset names, frequency, fill type, column (OHLC) and date range. We can add defaults. This is to allow different datasets ands parameters to be easily changed. These can be handled as API request parameters or in Python/R.
All reports must be well commented as in the report example attached.
I would like 3 simple reports to start:
a). Time Series Analysis on one time series displaying month averages with seasonality and moving averages etc.
b). As above but with analysis of relationship, movement, difference/correlation of THE 2 time series
c). Seasonality on one time series
Based on: https://machinelearningmastery.com/time-series-seasonality-with-python/
I would like you to explain about the reports you will create and what features each will have. My list above is only a suggestion. If you use another developers code we need to credit them in the notebooks. It also must be legal to use the code freely - even commercially.
You can use Python or R with Jupyter.
Docker experience very useful. You will assist in getting the notebooks installed and running on our server after they are created. We have Jupyterlabs installed with Python and R.
Thanks
Links:
https://machinelearningmastery.com/time-series-seasonality-with-python/
https://goodboychan.github.io/python/datacamp/time_series_analysis/machine_learning/2020/06/18/01-Predicting-Time-Series-Data.html
Related categories:
Python
Statistics
Machine Learning (ML)
R Programming Language
Statistical Analysis