Downloading
& Creating a Structured Database of Job Adverts
During
the last 2 months, Samata had been downloading from Naukri.
I
believe about 120,000 have been downloaded so far.
(We
also have 30,000 earlier set by Sajida.)
Occasionally,
Inder / myself helped download ads from Monster India, whenever we found
time.
Altogether,
in 60 days, we have barely managed to download 3,000 job ads/day (no
more).
There were problems of:
- Internet
connection bad / not at all
- Samata
had some other urgent work
Now
that we have settled down comfortably in our new office with a 512 kb line,
we must begin this job on priority & in right earnest!
Our
target should be 10,000 job ads/day.
This
will require 2 persons / 2 machines.
Vittal
should be one of them.
And,
if Deepa cannot be spared full-time for this work, I suggest you take any temporary
/ contract-based SSC boy for ₹1,500 p.m.
His
target must be 5,000/day (same as Vittal).
I
would like this arrangement to be implemented & stabilized by next weekend
— i.e., 21st Sept.
That is enough time to try a new PC (same configuration).
Of
course, clicking on a mouse is not human use of human beings, and should
be used only till such time that—
- We
get MYRIAD modified (at a price, of course), so that it runs 5
minutes every morning you instruct it — visiting job sites to download
automatically.
- OR
- Develop
a Spider on our own.
- OR
- Download
some spider freeware mentioned in Foster’s book (“Recruiting on the
Web” — now available with Kavita).
Maybe,
Abhi may wish to set a target for implementing such a spider — and free
Deepa/Vittal to do some human...
As
far as job-adverts are concerned, our first target is Indian job sites,
starting with
Naukri / Monster / JobsAhead / JobStreet / JobAds, etc.
Today’s
advert in Times of India speaks of 90,000 job-adverts available
on TimesJobs.com!
In
Google, if you type:
“India
+ job sites”
you
will get a list of 200+ job sites!
Our
spider should be able to download job-adverts from all of these.
Earlier,
we had tried Infogrists.com / Lexikot.com (?) for spiders (commercially
available) and did not succeed because at that time, our concentration was on
downloading resumes.
And
to access resume databases, all job sites require subscription + password.
But
this is not the case with job-adverts.
Job-Adverts
fall under the “Public” category & can be accessed / downloaded by
anybody without a password / subscription.
Gradually,
we must extend our efforts to USA-based job sites to download USA-based
jobs (IT jobs only).
This
is because all Indian IT professionals want to migrate to USA (except for those
working in 3P!).
If
we succeed in downloading 2.5 lakh job-adverts/month (@10,000/day × 25
days) and do this for next 4 months, then we have a large enough database for:
- Conventional
job search feature on Jobseeker.com (our next
website URL).
- Functional
/ Profile job-search.
(hand-drawn
curve showing “Match Theory Job-Search” with Match Index scale)
- Job
Alert as SMS on mobiles (1.7 million now, 10 million
by 2007).
- Job
Descriptions for GumAd module.
- Skill-level
keywords for self-learning software.
- Mailing
list of advertisers to promote RecruitGum.
- “Job-Market
Analysis” graphs (Trends)
- (Sample
given to Deepa 2/3 months back)
- →
For publishing in newspapers / websites etc. to give us LOTS of FREE
publicity!
- Delivering
JAWS (Jobs Across World Summary)
To
- Newspapers
- Cyber
Cafes
- Other
websites
- TV
channels etc.
→
On Jobseeker.com, there will be a “Configure Your JAWS” page
where subscribers will fill in & start getting JAWS at desired frequency.
- Creating
database of “Company-wise vacancies advertised” over the years.
- This
will help RecruitGum & 3P to market their services.
- Data-mining
at its best!
I
am sure you can think up of many other uses.
But
one thing is most important — whatever number of job-adverts get downloaded
during the daytime must be AUTO-CONVERTED & UPLOADED on Jobseeker.com
same night!
This process, too, must be automated.
There
should be no need for human intervention.
(signature
mark at bottom)
cc:
Sanjeev / Kartavya
No comments:
Post a Comment