WSO Company Database 3.0 - Thank You and Let's Keep Rolling
As we approach 1,000 companies in the WSO Company Database, I'd like to take this time to first thank everyone that has helped us get this far. The interns, the volunteers and the WSO community in general. We finally have a template that is workable and a critical mass that is meaningful.
I am super excited about this initiative and think this database could take the community to the next level. No more "Can you list the boutique banks in this region?" threads need to happen with people pulling up CapIQ lists. We have almost all of that (1,000+ companies) in the database now, and more is coming.
My hope and goal, is that over the next 2 years, we can do the same with the "Do you know how much X bank pays associates?" and "What is it like to work at XYZ firm?" and "What interview questions should I expect at XYZ firm?" type questions.
It is a monumental task but one I think we are capable of completing. In fact, a lot of that information is already in the over 500,000+ comments on WSO. It's just a question of getting buy-in from you guys and making sure we get the collective knowledge and insight of the group into a form that is easily searchable for the young chimps.
If we don't get 1,000s of reviews, interview insights and compensation data points, the Database will still be useful as a geographical search tool, but it won't reach it's full potential. I hope that the more experienced monkeys will join me in adding their past work history - company reviews, interview insight, and compensation data. We see this knowledge in the forums every day - I'm hoping we can get some of that in the database so that a true evaluation can be made of the firms on Wall Street. What firms are the boiler rooms? What firms pay great? What firms pay under market?
With that in mind, you'll see a new link under the Logo called "DATABASE"...you can click on it to take you to the DB, or you will see there are drop downs to take you straight to adding content.
Thank you for your continued contributions. Every day I'm excited to try and make WSO a better community - more fun and more educational. I hope 2012 will be a record year and I hope that database and all of you will help us get there.
Thank You,
Patrick
One thing that comments have
One thing that comments have that the database will not is the credibility associated with the poster. In all the comments/posts on here, some numbers/insights carry much more weight than others do, and it would be nice to carry that over to the database. Of course, this would be tough to do in light of anonymity concerns, though I suspect that you could come up with some "credibility score" algorithm that could anonymously give credibility to some additions and perhaps not others. For example, something like
credibility score = a + b + c + d + e
where
a=min(SB/MS,10) - min(MS/SB,10)
b=ln(|points|)
c=min((time since registration, in days/150) - 1,5)
d=if(certified,3,-3)
e=epsilon noise term to preserve anonymity, for example standard normal (0,1) term
Looks awesome, Patrick. Nice
Looks awesome, Patrick. Nice work.
[email protected]
Wall Street Mentors
Mock Finance Interviews
Dr Joe wrote: One thing that
One thing that comments have that the database will not is the credibility associated with the poster. In all the comments/posts on here, some numbers/insights carry much more weight than others do, and it would be nice to carry that over to the database. Of course, this would be tough to do in light of anonymity concerns, though I suspect that you could come up with some "credibility score" algorithm that could anonymously give credibility to some additions and perhaps not others. For example, something like
credibility score = a + b + c + d + e
where
a=min(SB/MS,10) - min(MS/SB,10)
b=ln(|points|)
c=min((time since registration, in days/150) - 1,5)
d=if(certified,3,-3)
e=epsilon noise term to preserve anonymity, for example standard normal (0,1) term
While this seems ridiculously complex, it's an interesting formula. I calculated my "credibility score" ignoring the epsilon noise term and came up with ~33. I then played with some of the variables (SB/MS ratio, bananas, time, etc.) and actually think the sensitivities are pretty decent. I'd probably work to modify (a) slightly so that new users with only a few SBs don't get a massive boost (perhaps say add SB/5 as another component of the formula). Dr. Joe, under your proposal, would a credibility score be assigned at the time of posting or continue to update as the user posts more?
All of that said, I think a credibility score is probably unnecessary. Just because a poster has been registered forever or posts frequently doesn't mean his/her compensation datapoint is any less accurate. In fact, I imagine to get any sort of scale out of the database we will be relying heavily on anonymous monkeys and infrequent posters.
CompBanker
Just a suggestion- Is there a
Just a suggestion-
Is there a way to solicit entries on certain companies? Say you have a salary question about Goldman Sachs- instead of putting up a thread asking people, is there a way to direct people to the company database (flag it somehow?). People with knowledge of the subject may be willing to reply to a thread but wouldn't necessarily take the initiative to fill out every bit of company info they can provide.
"Social cohesion and puritanical morality place roughly on my list of concerns between whether I'll pick up jock itch at the gym this week (not likely, since I don't go the gym) and whether it'll rain in Christchurch, New Zealand next Tuesday."
-Eddie
duffmt6 wrote: Just a
Just a suggestion-
Is there a way to solicit entries on certain companies? Say you have a salary question about Goldman Sachs- instead of putting up a thread asking people, is there a way to direct people to the company database (flag it somehow?). People with knowledge of the subject may be willing to reply to a thread but wouldn't necessarily take the initiative to fill out every bit of company info they can provide.
good question. right now, I have a few interns PMing certain users that display knowledge on particular companies (so it's obvious they were either an intern or a FT employee) -- besides that, I'm not sure there is an easy way to encourage additions.
Another option is to make the database available only once a user has added to it....the downside to this is that then we would have to block it for anonymous users and the search engines would not be able to crawl it either.
Right now, all options are on the table - the key here is to get as much information in the database as quickly as possible so that it becomes useful. I can get my team to keep adding companies, but I need the communiites help with Company Reviews, Interview Insights and Compensation Data.
Thanks,
Patrick
WSO Conference 2013
Private Certified User Chat
What about giving a free
What about giving a free guide to someone who gets certified and contributes all 3 "points" to the database?
Seems like that would build up the databases pretty fast while keeping the info credible. and it would add certified users to the site as well..
CompBanker wrote: Dr Joe
One thing that comments have that the database will not is the credibility associated with the poster. In all the comments/posts on here, some numbers/insights carry much more weight than others do, and it would be nice to carry that over to the database. Of course, this would be tough to do in light of anonymity concerns, though I suspect that you could come up with some "credibility score" algorithm that could anonymously give credibility to some additions and perhaps not others. For example, something like
credibility score = a + b + c + d + e
where
a=min(SB/MS,10) - min(MS/SB,10)
b=ln(|points|)
c=min((time since registration, in days/150) - 1,5)
d=if(certified,3,-3)
e=epsilon noise term to preserve anonymity, for example standard normal (0,1) term
While this seems ridiculously complex, it's an interesting formula. I calculated my "credibility score" ignoring the epsilon noise term and came up with ~33. I then played with some of the variables (SB/MS ratio, bananas, time, etc.) and actually think the sensitivities are pretty decent. I'd probably work to modify (a) slightly so that new users with only a few SBs don't get a massive boost (perhaps say add SB/5 as another component of the formula). Dr. Joe, under your proposal, would a credibility score be assigned at the time of posting or continue to update as the user posts more?
All of that said, I think a credibility score is probably unnecessary. Just because a poster has been registered forever or posts frequently doesn't mean his/her compensation datapoint is any less accurate. In fact, I imagine to get any sort of scale out of the database we will be relying heavily on anonymous monkeys and infrequent posters.
Practically, I would think scores should update based on after-post activity. Theoretically (and completely irrelevantly) this would create a privacy issue since (given one is bored enough) this could be traced back to user behavior.
Also, I think the "a" term should be changed to
a=min((SB+3)/(MS+3),10) - min((MS+3)/(SB+3),10)
which will address the issue of new users having huge ratios based on low denominators, as well as the issue of dividing by 0.
Lastly, I agree with your points about largely new users contributing majority of information, and that point makes this all a useless exercise. In fact, such a system could discourage new users from submitting data points since they would be treated as not credible. Oh well - fun exercise anyway. Could also be a pretty good solution to the previously posed question (in another thread) of ranking users in general by contribution - would be interesting to see the results. CB - think anyone could beat you based on this system? Seems unlikely since you max out all categories except points - perhaps evidence that the formula works.
Yeah, I think at this point,
Yeah, I think at this point, the key metric is # of data points, not some complex formula ranking the credibility of each addition.
Without a massive volume/ influx of Company Reviews, Interview Insight and Compensation Data, the database really just serves as a directory. So I'd like to encourage participation as much as possible...and I think our efforts (or my developers) is better spent on other projects rather than trying to put yet another ranking systme in place. what we have right now is pretty complex already. (WSO Credits, Banana points, silver bananas and monkey shits)...especially for the new guys.
anyway, i appreciate the brainstorming...thanks for your contributions so far guys. It is steadily growing because of users like you.
-Patrick
WSO Conference 2013
Private Certified User Chat
Dr Joe wrote: CB - think
CB - think anyone could beat you based on this system? Seems unlikely since you max out all categories except points - perhaps evidence that the formula works.
Patrick would be real close. His SB/MS ratio is about 8.1. Based on your formula, my ratio equation would only net me 10 despite my ratio being about 38.5. He has an extra year as a member (worth 2.3 extra points), but we would both cap out at 5 in that category. His higher banana count isn't worth many points at the levels we're talking about, so I'd probably end up about 2 above him.
CompBanker
...which would be fair since
...which would be fair since i think you've been more valuable with content contributions to the site. :)
WSO Conference 2013
Private Certified User Chat
hey guys, just wanted to
hey guys, just wanted to check in....are people having any issues filtering for foreign firms / offices?
Thanks
WSO Conference 2013
Private Certified User Chat
We now have well over 1,000
We now have well over 1,000 firms and counting....please help us fill in more Company REviews, Interview Insights and Compensation Data!
http://www.wallstreetoasis.com/wso-company-database
Thanks,
Patrick
WSO Conference 2013
Private Certified User Chat
Have taken a look through the
Have taken a look through the database the past few weeks, there's already some great stuff in there - looking forward to see it grow
Last I checked, when you
Last I checked, when you search a company using the regular search box, if a result comes up from the database (amongst other results from standard posts), it will say who authored it. I would be somewhat more willing to be open if I knew that loophole were sealed off.
wso1234 wrote: Last I
Last I checked, when you search a company using the regular search box, if a result comes up from the database (amongst other results from standard posts), it will say who authored it. I would be somewhat more willing to be open if I knew that loophole were sealed off.
very good point. sealing this off....we've made it so no company or reviews or comp data will show up in search until we figure out how to remove the author line for those results. that being said, cache will need to be cleared, so may take 24hrs before they all stop showing up in the general search. once we figure out how to get rid of the author line for the results, we will bring them back into the search.
thanks
WSO Conference 2013
Private Certified User Chat
Wow, one thousand companies!
keep it up guys...over 1,500
WSO Conference 2013
Private Certified User Chat