+ Reply to Thread
Results 1 to 5 of 5

Thread: Basic Estimation; Should I use future data?

  1. #1
    Points: 808, Level: 15
    Level completed: 8%, Points required for next Level: 92

    Posts
    11
    Thanks
    0
    Thanked 1 Time in 1 Post

    Basic Estimation; Should I use future data?




    Hi everyone,

    I'm new on the forum but had a question to ask that's related to the work I currently do.
    I'm wondering if we should include "future" or current data to estimate past events that we don't have information for?

    An example, If I count how many cars pass my house everyday for one month. But sometimes I miss a couple days here and there, but still want to estimate how many cars past my house when I wasn't able to be there. Should I be using the data that I have from future days than what I'm estimating? So if It's 8 Jan, and I want to estimate 5 Jan, should I just be using 1-4 Jan to estimate, or can I also include 6-8 Jan data to estimate how many cars passed on 5 Jan?

    Thanks

  2. #2
    TS Contributor
    Points: 12,227, Level: 72
    Level completed: 45%, Points required for next Level: 223
    rogojel's Avatar
    Location
    I work in Europe, live in Hungary
    Posts
    1,470
    Thanks
    160
    Thanked 332 Times in 312 Posts

    Re: Basic Estimation; Should I use future data?

    hi,
    this can be formulated as a missing value problem - and the answer is a cautious yes, as long as you can be reasonably sure that the process that generated the data stayed the same . E.g. Jan 4th is frequently "abnormal" because people might take it as a day off to prolong the new year vacation but Jan 6-8 are generally normal workdays so, obviously you could not use them to impute the missing data for the 4th.

    I think you should look at data imputation in general to see what possibilities you have.

    regards

  3. #3
    Points: 424, Level: 8
    Level completed: 48%, Points required for next Level: 26

    Location
    Oslo, Norway
    Posts
    14
    Thanks
    11
    Thanked 2 Times in 2 Posts

    Re: Basic Estimation; Should I use future data?

    Would agree with @rogojel.

    However, the example that you give us is trivial. If you could provide us with the actual problem at hand, or what you want to achieve, you will probably get more nuanced answers.

    Best

  4. #4
    Points: 808, Level: 15
    Level completed: 8%, Points required for next Level: 92

    Posts
    11
    Thanks
    0
    Thanked 1 Time in 1 Post

    Re: Basic Estimation; Should I use future data?

    Thanks very much guys.

    The problem I'm trying to solve is essentially this. If I'm sending out emails for a restaurant and I want to see how many people come in outside of and during an emailing period to compare them. But during the emailing period, I want to know how many people would have come in if we didn't send out those emails using an estimate to see if emails are actually assisting the business. This is a pretty basic problem, but one I'm trying to do on a somewhat constant basis.

    Do you guys have any advice? Should I just use what @rogojel said?

  5. #5
    Points: 808, Level: 15
    Level completed: 8%, Points required for next Level: 92

    Posts
    11
    Thanks
    0
    Thanked 1 Time in 1 Post

    Re: Basic Estimation; Should I use future data?


    Quote Originally Posted by factstalk View Post
    You could survey the people how did they learn about the restaurant or to compare the number of regular visitors with the number of customers after emailing.
    I've got the numbers to compare regular visitors with the number of customers after emailing. Would something like simple linear regression be sufficient? Or are there better methods out there?

+ Reply to Thread

           




Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts






Advertise on Talk Stats