Author Topic: Historical Data  (Read 1494 times)

kiltym

  • Newbie
  • *
  • Posts: 7
    • View Profile
Historical Data
« on: March 09, 2016, 12:13:34 PM »
Hi,

First, thanks for all the effort in this tool.  I am just starting to look into it and it seems very promising.

One question.  The Instant ROI and Projected ROI are based on ~380k notes.  It seems these were perhaps last loaded in Jan 2015.....?

Is there any plan to get this updated with more recent data?  The data I see on NSR seems "higher" for the same filter, and wondering if this is due to age of data, or calculated values.

Also, something to consider for the queries.  When I enter a min # of months for the last delinquency, say 36, the filter works, but it also excludes all those notes for which there is no delinquency on record at all.  I would think no delinquency should fall into a 36+ filter.  I understand why it doesn't, but an enhancement for the future perhaps.

Thanks again!
Mark

sociallender

  • Sr. Member
  • ****
  • Posts: 273
    • View Profile
    • Email
Re: Historical Data
« Reply #1 on: March 09, 2016, 08:19:34 PM »
Thanks for the feedback.  I am working on an updated version in my free time.  It will have updated historical data. The new version should have a single loan historical browser where it will show the ROI calculations.  There are more features that will be included as well but hard to find the time recently. 

I will also try to add a check box for the inclusion of NA for all fields.  May take some time but it will give a bit more control.  In the meantime, you could always create a filter manually.  An older version of PLS had this feature but never added it to PLS 7.  PLS 7 was a complete re-write so I had to determine what were the most important features.

Thanks again for the feedback and will keep everyone posted when I can get a new release out for testing.

gstrubinsky

  • Newbie
  • *
  • Posts: 6
    • View Profile
    • Email
Re: Historical Data
« Reply #2 on: April 23, 2016, 10:58:10 PM »
I loved the historical data since it would allow me to play with my filter parameters without finding out -the-expensive-way that the filter was no-good, but I don't seem to be able to find it.

The usual suspect would be the 'Completed Notes Only' in the filter definition screen. Then I re-load the data and there is no visible change. I downloaded the data provided by lending club into a database and it is VERY DIRTY! periods and commas are generously embedded into numbers so that the data can only be loaded as strings. It feels like someone typed the data manually in which does naturally make no sense with that amount of data. There will have been some manual work done on each loan, but the interest, the loan amount, the interest paid, should be represented as numeric value.

IF there is a way to explore different filter scenarios with clean(er) data and your system has those already, that would be grand. Either through your program's filters or exportable I would love to try 'playing' (since this is intended for REAL money investment, the term 'playing' is very loosely used) with different settings. This without having bound my -limited- financial resources  in trial-and- ERROR attempts would be absolutely grand!

Is there a way to get my virtual hands on this data through your system, and if there is, how would I do that. I see the problem of having millions of rows with a too wide filter on the historical data since 2007, but with a limit of returned rows or date limits or summarizing or in its simplest way refusal to deliver rows once the limit has been reached would probably work.

In any case, since it does not seem to return completed loans, what does  the  term   'Completed Notes Only' mean and most importantly can I see the success or failure in reaching a decent ROI through hind-sight = what-if scenarios = on historical data ?

with kind regards and very impressed
g strubinsky

Fred93

  • Hero Member
  • *****
  • Posts: 1896
    • View Profile
Re: Historical Data
« Reply #3 on: April 23, 2016, 11:14:50 PM »
I downloaded the data provided by lending club into a database and it is VERY DIRTY! periods and commas are generously embedded into numbers so that the data can only be loaded as strings. It feels like someone typed the data manually in which does naturally make no sense with that amount of data.

Which file did you download?