Skip to content

Python based web application to import, connect and analyze manufacturing data from multiple data sources.

License

Notifications You must be signed in to change notification settings

Omnind/AnalysisPlatform

 
 

Repository files navigation

Analysis Platform

2024-02-07: (Hotfix) Released version 4.5.2, which fixes the issue where graph display was completely disabled.
                     Please refer the "Upgrade" section to upgrade AP+DN7.

Analysis Platform is an open source web application to import, connect and visualize factory IoT data. It helps to collect, link and integrate data from multiple data sources. Visualizations include Digital Native QC7 Tools, which is designed especially suitable for data on manufacturing domain. Let's try data-driven process improvement by utilizing the data sleeping in the field.

FPP FPP PCP
CHM MSP PCP
AgP ScP PCA
COG SkD GL

What can we do with the Analysis Platform?

The main activities that user can take with the Analysis Platform are:

  • Check behavior of time-series or id-series data, stratify across processes (Full Points Plot)
  • Check behavior of fluctuation of the distribution (Ridgeline Plot)
  • Check correlation between variables and clusters in distribution (Multiple Scatter Plot)
  • Check behavior of each category/group (Stratified Plot)
  • Check behavior of related to human activities & process and product behavior (Calendar Heatmap)
  • Look for key variables that strongly relates to output (Sankey Diagram)
  • Grasp phenomenon that occurs at the same time (Cooccurrence Graph)
  • + many other useful visualizations

The Analysis Platform currently supports following data sources:

  • CSV/TSV/SSV(semicolon)
  • SQLite
  • MySQL
  • PostgreSQL
  • SQL Server
  • Oracle Database

Terms of Use

On your first access to the application, you must read and agree to Terms of Use shown on modal screen.
If you are going to use "start_ap.bat" to run the The Analysis Platform with the Windows embeddable package, running the batch file is regarded as you agreed to the Terms of Use.

Requirements

  • Windows (Tested with Windows 10)
  • Python (>=3.6) (Tested with Python 3.9.0, pip 22.2)
    • If you are going to use the Windows embeddable package, you do not have to install Python in your machine.
      (See section: For users: Run the Analysis Platform with the Windows embeddable package)

Getting Started

For detailed instruction, please refer below documents (PDF):

For users: Run the Analysis Platform with Windows embeddable package

If you use Windows machine, you can use Windows embeddable package to run the Analysis Platform without installing python. (See the document "Getting Started" for details)

You only have to download, unzip the file, and double click the batch file start_ap.bat.
The batch file will automatically download:

Note:
If you are connecting the internet using proxy server, 
you must edit "startup.ini" and specify the appropriate address in advance.

The Analysis Platform is activated after all downloads are finished.

The Analysis Platform uses Port 7770 by default.
Open below URL to access the application.

http://localhost:7770/

Downloads are only excecuted only if above files were not detected, so you can use use "start_ap.bat" for the next activation. (The Analysis Platform detects a folder named python_embedded_39 on the same level as AnalysisPlatform-master)

The Analysis Platform can run without internet connection. If you want to use this application on machine which has no internet connection (for example, intranet), you can first download all required files on other machine (which has internet connection), then copy entire files.

For developers: Run Analysis Platform with Python installed on your machine

The Analysis Platform uses Flask framework.

  • Python (>=3.6) (Tested with Python 3.9.0)

First, pull this repository to your PC.

git clone https://github.com/apdn7/AnalysisPlatform.git
cd AnalysisPlatform

Then, install requirements:

pip install -r requirements/common.txt

If pip install fails, try using Python 3.9.0.
Analysis Platform is activated by following command

python main.py

Corresponding ODBC driver must be installed to use SQL Server and Oracle Database.

How do we shut down Analysis Platform?

To shut down the Analysis Platform,
press shut down button on bottom of sidebar
(this button is only available on host machine, by accessing the application with localhost:{port_no}),

Upgrade

To upgrade the Analysis Platform to a new version, see below document:

Uninstall

To uninstall the Analysis Platform:

  1. Remove the Analysis Platform folder (the folder which the start_ap.bat is included)
  2. Remove the following files/folders:
    • get-pip.py
    • python_embedded_39
    • Oracle-Portable

Is there any sample data that we can use?

By default, the Analysis Platform contains sample data and corresponding settings to get an overview of each visualization.
Data is stored as TSV file under the subdirectories in /sample_data:

  • /assembly: Quality data
    • /1_parts_feed
      • 20220228.tsv
    • /2_inspection
      • 20220228.tsv
  • /parts_processing: Machine data
    • /1_machine_parameter_a
      • 20220311.tsv
    • /2_machine_parameter_b
      • 20220311.tsv
    • /3_finishing
      • 20220311.tsv
  • /COG_sample_data: Occurence of machine alarms
    • alarm_every_15minutes.tsv
    • alarm_daily.tsv
  • /AgP_sample_data: PartNo, Ok/NG, NG_Mode, and pressure data
    • AgP_sample.tsv

Above data will be automatically imported after activation.
You can call each sample visualization from 'Load' or 'Bookmark' on upper right corner of GUI.

If you do not need those data, you can either initialize the application by removing instance folder before the activation,
or you can just remove each data source setting from GUI.

License

The Analysis Platform is released under MIT License.
See our LICENSE for more detail.

Updates

You can check release notes for new features and bugfixes.

Troubleshooting

  • The Analysis Platforms fails to activate with an error message:
    ImportError: DLL load failed: The specified module could not be found.
    • You might need to install the "Microsoft Visual C++ runtime library".
    • Documents: EN / JP

About

Python based web application to import, connect and analyze manufacturing data from multiple data sources.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • JavaScript 64.4%
  • Python 23.9%
  • HTML 6.9%
  • CSS 4.5%
  • Batchfile 0.3%