Hidden Gems lets you identify relationships in your data using automated machine learning.

To get started, you only need two things:

  • You need a compatible CSV file containing your data
  • You need to know what question you want to ask your data

Simply upload your file and let Hidden Gems do the rest.

When Hidden Gems completes its analysis, it will provide you with a summary report, as well as generated models you can use to allow Hidden Gems to predict future data.

So what kinds of questions can Hidden Gems help answer?

Can I have everything on the menu? (Local Edition Only)

If you have data but don't know what you are looking for, this question is the right one for you.
It will evaluate every column in your data against both questions above, and if it can get a result it will let you know.

Is it or isn't it?

If you have data that fits into two general categories, this question is the right one for you.
Examples: Is a house going to sell at higher than asking price? Should I buy a particular car or not?
See a Sample File

What will the value be?

If you have data that all focuses around a single numeric value, this question is the right one for you.
Examples: What will a house actually sell for? What is the resale value of this car?
See a Sample File

After you run the analysis, you will be prompted to download your result. Here is what that ZIP file contains:

EvaluatedData.csv

If you included data where you don't know the answer to the question in the first column, this file will include only those records, but now with the answer filled in for you. If you are asking the question "Can I have everything on the menu?", you will get one of these for every column that was evaluated successfully.

GeneratedModel.hdgr

This file, the result of your analysis, represents everything that Hidden Gems learned from the original file you uploaded. You can use this later if you just want to evaluate new data but don't want to re-do the digging process. If you are asking the question "Can I have everything on the menu?", you will get one of these for every column that was analyzed successfully.

Summary.html

This file will tell you which of the columns in your file mattered most to the outcome. In other words, it will tell you which columns it thought was most important! It will also include raw statistics for your generated machine learning model in case you want to dig a little deeper.

  • If you like the app, please consider buying the Local Edition! [64-bit Windows and Linux Only] Hidden Gems Local Edition can run on your own workstation or server, has no imposed file size limit and lets you do an automatic comprehensive analysis of your data by asking the question "Can I have everything on the menu?".
  • It's best if you combine data where you know the answer and data where you don't know the answer. Hidden Gems will fill in the gaps, effectively 'doing your homework' for you.
  • If your data is bad, the results will be bad. What is good data?
    • First, good data is a LOT of data. The more the merrier.
    • Second, good data doesn't have a lot of missing pieces- most or all of your columns are filled in for each row.
    • Third, good data doesn't include dead-giveaway columns. So if you are asking what price a house will actually sell for, and if you include the seller's final offer, this app may not provide anything beyond obvious results.
  • This app is being provided as-is. Use at your own risk. Don't rely just on this to provide critical business decisions. It can be a helpful aid, but it isn't an end-all-be-all solution.