Data Scientist

Speciality Business / Secteur des papiers de spécialité Raleigh, North Carolina


Description

Who is Domtar?

We design, manufacture, market and distribute a wide variety of products including communication papers, specialty paper and packaging papers and absorbent hygiene products.  We operate the following business segments: Pulp and Paper and Personal Care.  We had revenues of $5.5 Billion in 2018; approximately 18% was from our Personal Care segment.  

                                            

Who is Domtar Personal Care?

Personal care is not just our business. It’s Personal. It’s a child, a loved one; it’s often about those that can’t help themselves. Caring for them can only be personal.  So Personal Care isn’t our category; it’s our calling, our mission.  We imagine the solutions that make caring for babies’ skin, and preserving dignity and independence easier and more affordable for everyone.  At Domtar, we believe everyone deserves personal care.

 

Our Vision

To be a global leader in absorbent hygiene by meeting consumers’ diverse needs through effective, affordable and widely available personal care solutions.

 

Our Mission

We champion health, dignity and comfort.

 

We serve consumers and caregivers wherever they access personal care, through healthcare, retail and direct channels.  We create value through our innovative solutions and achieve cost efficiency through operational scale, flexibility and the use of technology.

 

Your Role

The Data Scientist works with the business stakeholders to identify what the business requirements are and the expected outcome. Model and frame business scenarios that are meaningful and which impact on critical business processes and/or decisions.

 

Working with and alongside business analysts. This can be business requirements, such as reducing churn for mobile providers for example, or increasing cross-sell of products by suggesting other products of interest to the client. Identify what data is available and relevant, including internal and external data sources, leveraging new data collection processes such as smart meters and geo-location information or social media. Additionally, the cost of getting the data will impact on what sources should be considered, either because the data needs to be purchased from a third party or because of the complexity of the source. The data scientist will need to work with business subject matter experts to select what the relevant sources of information are.

 

Work with the data steward to ensure that the information used follows the compliance, access management and control policies of the organization and that it meets the qualification and assurance requirements of the business. Data stewards and data scientists will need to define the data quality expectation in the context of the specific use case.

Work in iterative processes with the business and validate findings. For example, validate discovered correlations across data with the business. Discovered findings can appear counter intuitive. Data scientists need to be able to expose the rationale to their findings in easy to understand terms for the business. Data scientists should be prepared to present back results that contradict common belief.

Include experimental design approaches to validate findings. Data scientists will need to validate analysis by comparing appropriate samples. In the case of clickstream analysis, it is easy to measure the impact of suggestions on total sales by splitting the users into two populations, one with and one without the recommendation engine. However, similar approaches to validation need to be identified regardless of the use case.

Suggest ongoing improvements to methods and algorithms that lead to findings, including new information. For example, refining suggestions based on clients’ interests, geographic location, age or gender, rather than just based on overall buying pattern.

Understand the use and ability to employ the appropriate algorithm to discover patterns. Examples include using statistical and data mining packages and dedicated frameworks such as Hadoop MapReduce.

Work with IT to support data collection, integration and retention requirements based on the input collected with the business. Examples here may include activities such as the definition of the blog collection process, definition and selection of the proper data infrastructure to perform the analysis, definition of the "perishability" of the data and the data archiving technology used to support the life cycle management of the data. The combination of the tools selected is likely to be very dependent on the use case and the assessment done along the 12 dimensions of big data/extreme information.

Educate the organization both from IT and the business perspectives on these new approaches, such as testing hypotheses and statistical validation of results. Given how new these approaches may seem too many people, it may come across as esoteric. Helping organizations understand the principles and the math behind it will be essential to driving organizational buy-in.

 

METRICS FOR EVALUATION:

Improvement to the business metrics for the overall project. For example, if the objective of the initiative is to improve cross-selling, the suggestion engine offering clients other, relevant products to buy should positively affect the cross-sell key performance indicator. As there are often multiple iterations of the analysis, the contribution to the improvement should be monitored initially and over multiple iterations.

Demonstration of the following scientist qualities:Clarity, Accuracy, Precision, Relevance, Depth, Breadth, Logic, Significance, & Fairness.

This is particularly important when presenting back the results of the analysis to peers and business subject matter experts.

Business user satisfaction. The ability for data scientists to communicate and work with business subject matter experts is essential.

Timely delivery of analysis. Due to the experimental approach to the problem, data scientists are at risk of over analyzing. Being able to define a "good enough" result is critical. Further refinements will happen over multiple iterations.


GENERAL RESPONSIBILITIES

All employees are held accountable for the following responsibilities which are measured in performance reviews:

 

Safety – Understands the Standard Operating Procedures for the job and performs the job in a safe manner.

 

Performance Standards/Quantity of Work – Meets the performance standards established for the job.

 

Quality – Meets the quality standards for the position.

 

Initiative/Work Ethic – Looks for ways to make things better.

 

Cooperation/Teamwork – Shares information and works together with others to promote, support, and achieve established goals. Follows directives as assigned by management.

 

Knowledge of Job- Understands the principles, concepts, techniques and requirements necessary to accomplish the job duties and to comply with all Company policies and procedures.

 

Problem Solving/Decision Making – Identifies problems and recommends potential solutions.

 

Attendance/Dependability- Understands the importance of being at work on time and on a regular basis.   

 


SPECIFIC RESPONSIBILITIES

% OF TIME PERFORMING TASK

 

 

Elicit requirements using interviews, document analysis, requirements workshops, surveys, site visits, business process descriptions, use cases, scenarios, business analysis, and task and workflow analysis.

 

Working with and alongside business analysts. This can be business requirements, such as reducing churn for mobile providers for example, or increasing cross-sell of products by suggesting other products of interest to the client. Identify what data is available and relevant, including internal and external data sources, leveraging new data collection processes such as smart meters and geo-location information or social media.

 

Proactively communicate and collaborate with external and internal customers to analyze information needs and functional requirements and deliver the following artifacts as needed: (Functional Requirements (Business Requirements Document), iii. Use Cases, GUI, Screen and Interface designs) Assessing whether data is fit for use by performing initial validation of data delivered as part of the project.

 

 

10-20%

                             

 

10-20%

 

                                                                    

 

 

 

20-30%

Include experimental design approaches to validate findings. Data scientists will need to validate analysis by comparing appropriate samples. Work with IT to support data collection, integration and retention requirements based on the input collected with the business.

 

Successfully engage in multiple initiatives simultaneously.

 

20-40%

 

10%

 

 

Participate and lead in training opportunities as well as conduct self-lead training over relevant topics.

 

 

Suggest ongoing improvements to methods and algorithms that lead to findings, including new information. For example, refining suggestions based on clients’ interests, geographic location, age or gender, rather than just based on overall buying pattern.  Understand the use and ability to employ the appropriate algorithm to discover patterns. Examples include using statistical and data mining packages and dedicated frameworks such as Hadoop MapReduce.

 

20%-30%

 

10%-15%