search

sfu-db / dataprep

Open-source low code data preparation library in python. Collect, clean and visualization your data in python with a few lines of code.
http://dataprep.ai
MIT License
2.08k stars 206 forks source link

Time Series Analysis #485

Open dylanzxc opened 3 years ago

dylanzxc commented 3 years ago

Is your feature request related to a problem? Please describe. The goal of this feature is to add more useful visualizations, statistics and insights to plot(df, x, y) for time series analysis.

Describe the solution you'd like I propose the following functions and tabs:

  1. Detect if the data has regular time intervals. If not, give transformation suggestions or provide transformation functions. More researches are needed. Estimated timeline: Jan 21 to Jan 27
  2. Basic Visualization: Line chart and Box plot(edit on the existing functions) In the line chart instead of auto-grouping to 15 groups , we plot all the data points since the dataset here is in time-series format. But we keep the time_unit parameters for line chart and box plot. image image

Estimated timeline: Jan 27 to Jan 29

  1. Visualize Seasonality: Multi-Line chart and box plot image image

Estimated timeline: Feb 1 to Feb 9

  1. Decompose times series Additively and Multiplicatively image

Estimated timeline: Feb 9 to Feb 16

  1. Autocorrealtion function(ACF) and Partial autocorrelation function(PACF) image

Estimated timeline: Feb 17 to Feb 24

  1. Lag Plot image

Estimated timeline: Feb 25 to Mar 3

  1. Test Stationarity: Unit Root test Augmented Dickey Fuller test (ADF Test) Kwiatkowski-Phillips-Schmidt-Shin – KPSS test (trend stationary) Philips Perron test (PP Test)

Estimated timeline: Mar 3 to Mar 5

  1. Estimate forecasibility: Approximate Entropy

Estimated timeline: Mar 8 to Mar 10

  1. CHTest: determine if seasonal differencing is required to stationarize the series

Estimated timeline: Mar 10 to March 12

Describe alternatives you've considered Dealing with missing values: forward fill, backward fill, linear fill, cubic fill, KNN mean, seasonal mean

Additional context N/A

jinglinpeng commented 3 years ago

Reference: https://traces.readthedocs.io/en/latest/ https://joachim-gassen.github.io/ExPanDaR/