This question may seem to trivial but without a sufficient background in mathematics i get to ponder how i could achieve what i am going to explain now.

If i have a supermarket for example and i get not uniform sales:thus unstable :

2/19/2014 : 12340 $
2/20/2014 : 1200 $
2/21/2014 : 556343 $
2/22/2014 : 54340 $
2/23/2014 : 43340 $
2/24/2014 : 531340 $
2/25/2014 : 50320 $

Assuming i want to know the growth rate of this supermarket based on the daily income, how may i archive this ? I have seen some examples but the income used is increasingly uniform.


The first step should always be, plot the data. Just do a simple scatter plot and look at it. See if you can "visualize" a line passing through the data in some sensible way. You might even print out the plot and try drawing a straight line through the "cloud" of data points with a ruler. If you do that you can figue out the slope of the line and have some kind of very rough estimate of a growth rate.

If however, when you plot that data you see what looks like a shotgun blast of data points, with no obvious trend, there isn't really much you can do...well, there isn't much you can say if you are being honest. Some will still fit a line to the data that says what they want it to say...(see how to lie with statistics).

Now, if there is a trend and the spread of the data are not too large, you might try to fit a line using simple linear regression. You might also try a simple moving average. Create a new data set where each point in the new data set is the average of the last k days (you choose the value of k). Again, plot the new data for various values of k to see if it looks like there might be a trend.

However, Sales revenues often have a seasonality. Sales on Sundays might always be significantly higher than on week days for example. A sales at certain times of the year will be higher than in other times of the year and so on. If your data have a seasonal component, you may like to try to "subtract it out". This is done by creating a new data set where each data point in the new data set is the difference between two data points in the original data set. Taking the "sales are higher on Sundays" example, if $x_i$ is the $i^th$ data point in the original data set, construct a new data set such that $y_i = x_{i+k} - x_i$ (k= 7 in our example). The parameter, k is called the lag.

Anyway, try these things first. Just using simple scatter plots and see what your instinct tell you.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.