Predictive analytics for Sales

I need to do predictive analysis for my products and need help with that.
I have a brand under which i have multiple products (ex P1, P2, P3, …Pn) and there are dealers who sell them. how can I use KNIME analytics platform to do a sales prediction for my various products. i have data from last 5 years.

get some inspirations from the Example server and the learning section

I can think of a few examples that might show you the possible directions

This site has an introduction to model building with KNIME and uses an example that could be similar to your question:

Concerning the predictions you want you will have to think about how to set up the analysis and prediction. Would you make a prediction for every single product and dealer and what role does the time play in your analysis. Sales within a month or a week.

Then the quality of a model will very much depend on the data you have and if there are informations in them that lead to a pattern to predict sales (like holidays, demographics around the stores). And typically sales can be influenced by several factors like advertisement or (planned) sales campaigns, discounts etc. You might want to think how to include these factors in you model, and if you could reproduce them for future predictions.

And you would want to establish a measure that gives you and idea about the quality of your models. A typical statistic would be RMSE

The KNIME Numeric Scorer has you covered there:

Also a correlation coefficient might be good (Pearson or Spearman).

And the KNIME Learning Hub has a lot of useful informations about data mining and model building:
https://www.knime.com/learning-hub
(under Applications / Data Mining)

1 Like

what is advantage of KNIME over Talend for ETL processing ?

I cannot speak about Talend since I have never used it. I think it depends on what you want to do and into which environment you will have to integrate and what tasks you have and what kind of people and talents you have access to.

My experience with KNIME is in these stages

  • standalone on a Mac/Windows machine (limited to what power this machine brings) - you might connect it to a wide range of databases and sources
  • KNIME Server ($) gives you power of automation and team collaboration
  • KNIME connection to Cloudera Big Data cluster where KNIME is you GUI/cockpit and automation engine

The philosophy of KNIME I think is well captured in the first article below. Give all analytics people in a company an affordable, easy to use graphic software for all kinds of tasks, a tool that would happily connect to all sorts of data sources and can run code from R and Python on top.

Sometimes hardcore coders do not like KNIME that much because they say: I can do that in xy lines in Python. Yes they can, but KNIME would also do that and one great benefit from my perspective is it sort of automatically does its own documentation thru the nodes and workflows. If you put in just a little more effort you can create a documentation on top.

So if you want to empower a broad range of people to use all kinds of analytics and automate some tasks easier than with VBA or some programming framework then KNIME is your thing especially if you can create a community around it that share code and examples (like on this forum but you might also encourage that within your company).

Considering ETL I see these possibilities in KNIME

  • use the integrated KNIME nodes for data manipulation (everything will be done within KNIME)
  • use the Database connections with generic nodes that require little or no knowledge of SQL
  • use KNIME as a GUI/front end for your own SQL jobs and send them from an workflow to your data base/big data environment that does the heavy lifting

Here you can see a small example how such an ETL job might work (discussion about SQLite that shows the potential of SQL-connections from KNIME)


Empower Your Own Experts! Continental Wins the Digital Leader Award

Gartner 2018 Magic Quadrant for Data Science and Machine Learning Platforms

Great Power, Great Responsibility: The 2018 Big Data & AI Landscape

1 Like

what is advantage of KNIME over Talend for ETL processing ?

The advantage is that KNIME is not limited to ETL. Given that Talend also offers an open source variety, you can easily compare the tools for yourself on the ETL aspect.

1 Like