1. Can Data Science
become a game-changer
in talent management?
Case Study
Experiment
by ELEKS Data Science team
2. Technological innovation
driven by professionals
Nowadays organizations gain competitive advantage with technological innovation
that includes wearables, smart analytics, digital transformation and IoT strategies.
Technological innovation is driven by professionals, from developers
to testers to product managers and CTO’s.
The way to outrun your competitors is to hire more talented staff
and develop existing employees in the most efficient way.
3. So we’ve asked ourselves a question here at ELEKS Labs:
Can we somehow use data
to improve the way
organizations hire,
manage and
develop their talents?
4. Challenge
Create a model that “understands” talent behavior,
perspectives, possible career development paths and influence factors,
and helps organizations make better talent management decisions.
5. Solution
A tool that is capable of predicting
possible career development paths for users
based on their previous experience, industries they worked in,
skills, education and more.
But, more importantly, this model suggests which skills,
education or career steps are required to be more successful
and reach one’s career goal in minimum time.
CareerWhiz: A Data-Driven Tool for Career Advice
6. Prior to building a model we have collected and processed publicly available data:
We collected over 11 mln publicly available vacancies,
resumes, recruiting requirements, job and position descriptions
and much more.
Raw data, containing all information was over 145 GB in size
We used natural language processing computer techniques
to filter all the text data
Analyzing this dataset helped us create a database of 80 000 positions of
various seniority levels from 100 industries with over 17000 parameters
describing each position.
Solution
7. Solution
As a next step ELEKS mathematicians and data scientists
built a model that can predict possible career development
paths based on user profile.
Building such a model means creating a mathematical function
that analyzes user profile as an input data
and provides prediction as an output.
8. Solution
To give user a personalized prediction we collect data from his LinkedIn profile,
split and transform it into a dataset, acceptable by our model
and use it as an input file for the model.
The model has over 17 000 parameters to describe each position and each user.
The more complete user’s LinkedIn profile is the better we understand
what type of professional he is, and the better is the accuracy of predictions.
To ensure a more precise recommendations for each user
the model utilizes collaborative filtering technique.
User data collection and preparation
9. Solution
Utilizing specially developed mathematical function
we predict potential career paths for each user.
We present 3 most probable results based on user’s education,
job experience, completed projects, summary, geography and more.
We analyze thousands of records in our DB in real-time to identify ones
with similar skill sets and assign current user to appropriate cluster.
And then make recommendations taking into account
specific skills from that cluster.
Recommendations include skills, education, job types profiles,
projects that are required to reach one’s goal.
Recommendations may even suggest changing
geographical location to better succeed in one’s career.
Prediction
10. Sample results
Olivia is a research intern at educational foundation, mainly involved
in research, marketing and communication activities,
which is stated on her LinkedIn profile.
CareerWhiz shows 3 potential career development paths for Olivia:
To become an executive in marketing Olivia needs to work
for another 5-7 years potentially as marketing coordinator,
project manager at FMCG or telecom organizations,
master and improve 11 new skills, including management, project management,
marketing, advertising, public speaking and more.
Olivia also needs to acquire new education, including MBA.
Executive in Marketing
Consultant at Education management organization
Co-founder, CEO at IT organization
11. Click on a picture below to access CareerWhiz and get your recommendations (careerwhiz.
eleks.com):
See how your career can evolve
12. What started
as experiment...
And it worked!
CareerWhiz: A Data-Driven Tool for Career Advice
is a useful and a smart tool that can help
many people reach better results in their career.
But this is still a proof of concept, that was built to prove
our hypothesis that data can help in managing talents…
13. Business benefits
We are working on developing this model further to bring real value to business
and here are some examples on how we see this can be done:
Improving quality of employees education
by understanding which skills and knowledge
are leading to better results for specific
professions and positions
Improving the hiring and promotions decisions
by understanding which employee
will perform better at specific roles and positions
Building internal career plans for employees
that will both guaranty higher motivation
and job interest for employees and
better results and productivity for employers
Identifying factors
that lead to success as well
as leaders capabilities among employees
14. Join us
... and explore how your organization can benefit from
using data for talent management and bring technological innovation
to daily operations.
15. About ELEKS
Named a Top 100 Global Outsourcing Company, ELEKS is a global organization
providing software engineering, technology consulting
and quality assurance services.
Since 1991, ELEKS innovative and award-winning solutions have significantly
contributed to the customers’ unparalleled business growth to include
Data Science, Mobility, Digital and Financial solutions.
Contact us
Eleks, Ltd.
7 Naukova St., Building G
Lviv 79060, Ukraine
phone: +380 32 297-1251
fax: +380 32 244-7002
ELEKS Software UK, Ltd.
5 Harbour Exchange
South Quay
London, E14 9GE
phone: +44 203 318-1274
ELEKS Headquarters UK Office
Find us at eleks.com
Have a question? Write to eleksinfo@eleks.com