Orange free Data Mining Tool

I was looking through 101-useful-websites article and came across AlternativeTo.net and used it to look up alternatives to say “Revit” and “AutoCad” and other tools I use.

I then typed in KNIME which I use for data mining, data analysis and it came up with Orange as a free alternative.  So I looked at some videos on it on YouTube

The point that drew my attention to the programme straight away was that it seems to have an SQL Database link (unfortunately I am having difficulty getting python pymssql to load which it needs) and also the quick visualisation with simple scatterplots and reviewing data quickly with datatables. The other nice thing is you can use Python Scripts.

Testing

I filtered data using Open Refine & Excel to simplify room data from a properties file, looking at room sizes for room types. The table is over 18000 rows long.

Item 1. Import CSV file

Item 2. Filter Rows for only displaying Floor Finishes (for getting floor area) and

Item 3. Select Room Names & Quantity (Floor Area. m2). 

Item 4. Use Scatter Plot to show X axis= Room Names &  Y axis = Quantity (Floor Area. m2). 

On initial testing a few things arise.

• You can create reports that then can be exported as HTML files, so you can post to web.
• Maps are static PNG files, so if you want to zoom in you have to save a few files to show different magnifications
• Graphs/Tables can be saved in the Reports and also as PNG files too.

End Comments

This is pretty fast with the excel files. It processes the data and displays it quickly.

I would like to get the Database connections working. The MS SQL is being a bit difficult in Python so I may try and use the PostGres connection and try it with the OpenMaint Database.

I will definitely be giving this a better run. It has tools that I am interested in and I’m keen to try using the Python Script & the Text Mining Tool.

There is a machine learning tool that looks as if I could stumble my way through it and I intend to give that a try too.