What percent of the active users on talk are the active membership at the space. Or some graph that shows the relationship
2 Likes
Excellent question. Admittedly I honestly thought you were aiming to get those numbers when you wrote “I’m working on getting the DMS member data, ratios and charts …”
(I literally copied and pasted the above 2-column text data into a text file and used that as the source data for the average calculation. If I have access to this data and other Talk related data, I don’t know about it.)
Access can be made … we have the power …
We have things like Door Swipes to DMS Username …
We have things like Talk logins …
We don’t currently have things like Talk logins to DMS Usernames … in the plans
2 Likes
How did you calculate that?
If they haven’t logged in I don’t think they are tracked. SO anyone just reading probably doesn’t show up in the count.
There are a whole lot of people in the list that it shows that visited zero times in the past month.
Does someone with no ID doesn’t appear in the count? I.e. Someone randomly looks in, no account, do they show up in count? If so, how are they differentiated as unique views?
It is only listing visits from user accounts on the system, if they are logged in.
It is not listing visits from unknown people or not logged in people.
I’m not certain if I am explaining it correctly. It can’t know who a person is, if they are not logged in, so it doesn’t count those people in this data.
1 Like
I managed to get more data out of the system and I’m not certain how to take that data and get an average days per logged in user.
Does this seem right … take the two columns and multiply them and create a third column …
and add up the user count column to get a total user count and the new column
and divide the new column by the total number of users?
In general, I use Octave/Matlab for all things mathematics related.
I used the built-in command trapz() to integrate the area underneath the curve seen in your first graph (Days visited vs Users) and then divided by the range of the data (which by definition is the average).
And depending upon your level of interest, you may duplicate the work …
[spoiler]
>>load data; %"data" is the text file containing the x/y values I found on Talk
%the first column is the number of days visited
%the second column is the number of users for a given number
%of days visited
% The x-data is a cumulative sum of the 2nd column of data
% which is what is seen in Draco's first plot
% /\
% /----- -------\
>> trapz(cumsum(data(:,2)), data(:,1)) %this gives area under a curve
ans = 5535.5
>> 2552-1841 %this is the range of the data
ans = 711
>> 5535/711 %this is the average (area under curve/range)
ans = 7.7848
%This is the same as above except the first data
%point is excluded in the calculation
>> trapz(cumsum(data(2:end,2)), data(2:end,1)) %this gives area under a curve
ans = 5444
>> 2552-2024 % this is the range of the data
ans = 528
>> 5444/528 %this is the average (area under curve/range)
ans = 10.311
[/spoiler]
1 Like
It is interesting but there is quite a bit of it hidden nor do I understand …
Actually, I’m wanting something I can put into Excel or some non-math orientated programming language like php or python.
I realize Python is fairly generic and Octave/Matlab somewhat more specific. However, truth be told, if one were to use Python, I’d recommend they use the Numpy library rather than writing ones own cumulative sum function and integral function that uses the trapezoidal method.
From the Numpy library, the two following built-in functions exist …
https://docs.scipy.org/doc/numpy/reference/generated/numpy.trapz.html
https://docs.scipy.org/doc/numpy/reference/generated/numpy.cumsum.html
Of course, then you’d end up with a script that looks striking similar to the one I used in Octave.
1 Like
Pandas (https://pandas.pydata.org/) is handy for this type of thing too.
I rarely log out of Talk, how does that show
1 Like
Ditto. And I’m logged in on multiple devices.
1 Like
I’m logged in on my IPad & on my Iphone.
1 Like
How does Talk “know” a user is a DMS member? Is it done via the email address? Could link membership to talk membership that way if so.
1 Like
Keep in mind, talk doesn’t auto-recognize if the account is a DMS member. So there is an amount (who knows how large) of DMS member accounts who have not been added to the DMS Members group.
2 Likes
Does this change in the near future?
With the majority of Talk locked-down to Green Dot users, I would think moving forward, an extremely good/accurate guess could be made. That said, I don’t know what data is accessible behind the scenes.
1 Like