Contents
The first step towards error calculation is to collect the data. Usually the values will be measured by hand, so one will go to /index.html and add a new row.
A row represents one series of measurements, for instance if one has to measure the commute of a pendulum, he will do so about 5-10 times for every parameter. So you have to enter the number of measurements in the box before the "Add Row" button. Then click on "Add Row".
If you have finished one series of measurements, click on "Add Row" again.
If all the samples are inserted the errors can be calculated by clicking on "Calculate". The button "Download" will open a new tab with the calculated results.
By providing and "x" value for each row you allow the server to create a new plot. It uses the matplotlib library to generate a scatter plot displaying all your values and a simple linear scaled plot displaying average with errorsbars, min, max and modus.
This plot cannot be used for further research. It is just created to give one a hint how the data looks like and is therefore plotted using the xkcd style.
The download function supports currently three formats: csv, json and ljson. These should fit all purposes.
The calculation is based on the following formulas:
Modus:
>>> Counter(samples).most_common()[0][0]