Lee Baker - CEO Chi-Squared Innovations

Hi there,

I have been a data scientist for many years, working mainly in the medical and academic arenas, and I had become very frustrated by how long it took to get the results that I needed from the raw data. It can take weeks to clean the entry errors and pre-process the data before analysis. Selecting the most appropriate statistical measures from standard statistical packages and knowing how to use them to reach the ‘story’ of the data can also be a real challenge.

Over the years some of my colleagues even admitted to "clicking buttons at random until I get an answer"!

Oh dear...

And what help do we get from current statistical packages?

  • What happens if you use an inappropriate statistical test on your data? Will SPSS tell you?
  • Will it tell you that the answer is - at best - inaccurate, or just plain wrong?
  • What happens if you choose the most appropriate statistical test and get the correct answer, but interpret the result incorrectly?
  • Will it tell you? Will it care?

These were some of the little things that troubled me about the current commercial statistics packages, but there were other things - bigger things - that troubled me more. When you have multiple analyses to do - many pieces of the jigsaw puzzle - how do they all fit together? Will your stats package tell you or will you need to ‘hit the books’ for weeks or months to find out? And will it help you to run multiple analyses or will you have to do it manually?

What happens when your dataset is large and there simply isn't enough time to run all the analyses with your manual stats program?

  • Should you ‘cherry-pick’ only those analyses that you think are useful?
  • Will you worry that you might have missed something important?
  • Are you seeing the ‘bigger picture’?

And what happens when you have spent weeks and months doing the analyses and you have all the results at your fingertips? Can you visualise how everything fits together? Can you explain it to someone else? Can you explain it all in your paper or thesis, or will you need to simplify and choose only those results that are central to your hypothesis?

In my experience, less than 10% of any dataset is analysed, and typically only 1% of the results are published.

Is this really the bigger picture?

Is this really what you want from your stats program?

To put it mildly, I was really unhappy with statistics packages - they seemed to make a difficult situation even more difficult

So I set about a plan to fix it...

Our vision is to produce statistical analysis systems that take you ‘from data to story’ with as little fuss as possible. To do this they should include:

  • full automation, from beginning to end
  • one click data cleansing
  • data pre-processing
  • correct selection of the most appropriate statistical measures
  • just what you need - no more, no less
  • explanation of the choice of statistical measures, so you can ‘learn as you go’
  • easy to understand interpretation of all the results
  • simple language not complex statistical terms
  • an intuitive visualisation of how all the results fit together
  • an explanation of how all the results fit together to form the bigger picture
  • the facility for real time collaboration with partners anywhere in the world

Most of all, to be truly valuable to your research it should do all this in minutes rather than months, and this is what we do.

Why not check out our products and services?

We understand that the story of your data is valuable, and we are determined to present that story to you with efficiency, practicality and absolute clarity...

So tell us your problems. What would make your life easier? You never know, between us we might just come up with something extraordinary...

Your success is our success, and by combining excellence, honesty and innovation we strive to ensure that you get the correct result - first time - every time

We love working with our new products, and we hope you will too.

I look forward to hearing from you

Lee Baker
Co-Founder & CEO
Chi-Squared Innovations