Data Analysis Skill Set

Values

These might vary with company culture, but are fairly common:

Company Culture

• Customer Focus - In a product that provides a service, its all about the end customer, you need be helping the end customer in a useful way, don’t hurt the view that the customer has on you.
• Take Ownership - Its your baby, everything about it, is on you, make sure the project goes all the way through to completion.
• Sense of Urgency - Move in an agile way, plan, monitor, re-think, but move onwards at a good pace.
• Highest Standards - Produce work you are proud of.
• Humility - Everybody makes mistakes, everybody has different views and these augment an idea into something bigger, listen to others.
• Exposure points - think what part of your work gets exposed to others, that helps prioritize what is important and what really makes a contribution.

Data Analyst

• What is the truth? - The job is to reveal the truth, its OK to have an opinion but do be on the lookout of biases influencing you(from you or others).
• Macro picture thinking - analyst needs to constantly asking “so what?”, what does this really means to the bottom line

Data Analyst ‘Hard’ Skills

Data Collection and Manipulation

• Basic ETL developer: data collecting, scraping from online sources, augment collected data with useful dimensions (often the time dimension).
• Telemetry: define adequate telemetry to tackle a problem, avoid over-complicating, but guarantee we will get all data needed for project.
• Data manipulation: table data (pivot, unpivot), csv, Json, XML.
• Basic understanding of dimensions and facts in a data warehouse and data-stores in general.
• Relational databases and SQL querying.
• Big data tools: hadoop, hive.
• Python

Data Analysis Methods

• Data QA: validating data to make sure is correct, try to understand data (http://bit.ly/1CWmrNc).
• Back of the envelope calculations.
• Define KPIs: understanding what is important, the organization context and domain, understanding the data being captured (how is captured, what limitations it has), surfacing what metrics more closely relate to the data analysis challenge at hand.
• Understanding A/B tests and experiments: understands the process on how to test: sample size calculation, calculate confidence in results.
• Understand Clickstream data (and web analytics data as a sub type of Clickstream): users, sessions and hits and their scope. Understanding why Month unique users is not same as the sum of 4 weeks of unique users.

Visualization and Sharing

• Visualization: understanding what charts to use, both when analyzing and presenting/sharing data.
• Regular Reports: what to include, what not to include, how to best display findings in a fair summarized format, how to automate regular reports ( or minimize update effort ).
• Presenting / Sharing Data: putting a slide deck together that tells the story of the findings, adequate summaries and adequate charts.
• Infographics
• Excel
• Understanding of chart types: Time series, funnels, Maps, heatmap,
• Interactive, creative charts: D3.js, Tableau, MS Power BI stack

Modeling / Statistics

• Statistical Methods: median, correlation, what methods to use in which situations.
• Prediction methods: regression, moving days average.
• Data modeling (and Machine learning): when to use, what they are good for, how to use.
• Hypothesis testing: p-value, Chi-Squared Test, T-Test
• R

Data Analyst ‘Soft’ Skills

Analytics Skills

• Understand the impact that the analysis results will have.
  • Understand well the problem statement.
  • Understand the context / domain / business.
• Tackle an analysis problem in a structured way, understand dependencies and limitations upfront.
• Includes reliable QA methods in the analysis process.
• Seek for unbiased truth. Reveal the reality, shielded from personal or stakeholders biases.
• Understand and be skilled with the analysis toolkit.
  • Query and manipulate data.
  • Know how and when to apply adequate statistical methods.
  • Know what is best chart to use for each type of data.
  • Good understanding of tooling available and how to use it.
• Clear communications: make sure the results are well communicated:
  • In document form: clearly written, easy to understand for the intended audience, adequate content, adequate charts, not sloppy, etc…
  • In presentation form: compelling presentations for any level audience.
• Takes ownership of project, proactive attitude: drive for resolution, impact analysis, pro-actively alerts who needs to be aware of impacts, make sure results are consumable by who needs them (and on time).
• Understand, follow and advocate the data privacy policies.
• Know how to prepare adequate telemetry, that collect the results that answer the projects needs.

Effective Work

• Can understand the problem and priorities, and self-sufficient to make a plan, and set priorities that are optimal to drive the project, understand which projects/tasks are more important and why. Example: how big impact can create, is it running now in production and could be impacting users (urgent), etc…
• Identifies and highlights risks early in project, and plans for them.
• Understanding of how current work impacts the overall org./biz
• Can influence/drive progress with articulating the value of the project. Project advocate.
• Can give time estimates and can articulate areas of risk, dependency and confidence in relation to ETAs.
• Tracks and documents project progress on a regular basis.

Communication

• Understand the request clearly, actively investigate numbers and probe further all parties involved to make sure the request, task and assumptions are clear.
  • Summarize and translate a problem statement into a few bullet points and confirm with stakeholders that it is correct - this helps confirming understanding, and agree on a clear path to tackle them.
  • Remember: assumption is the mother of all mess up’s. Ask, re-ask, to confirm assumptions.
• Agree with stakeholders the task requirements early in project.
  • Set expectations
• Keep communicating regular as the project develops.
  • Keep track and agreement w/ stakeholders, and make it transparent of goals, objectives, next steps, milestones, etc…
• Create good written summaries of the results, including explaining well complex data and choosing adequate charts and data explaining methods for it.
• Masterful storytelling: clear, contextualized and engaging outcome-oriented storytelling(written and presentations).
  • Can influence and give convincing verbal communications.
  • Can articulate well complex technical details.

Stakeholder Partnership

• Well defined responsibilities, accountabilities. What is expected.
  • Share with stakeholders the analytics plan and get their approval.
  • Deliver on plan.
• Manage expectations by communication well and checking of understanding.
• Flag up dependencies, risks and offer solutions (next steps) to problems.
• Look at the stakeholder problem/project from ground up (extensively) and suggest KPIs, reports and frequency that best tackle the problem.
  • Ideally needs to be included in the planning discussions (and in defining analysis solutions) in the stakeholder project sessions. (but depends whether stakeholder is by default the data driven or not).
• Follow and respect the agreed processes when requesting time and work from other teams:
  • Create a task in a ticket system, requesting the work.
  • Book a meeting when hard to reach someone by mail or IM.
  • Etc…

Development

• Look to increase visibility of the work done, share results widely, monthly/weekly summaries.
• Have a personal development plan: be aware of personal development needs, and create a plan to tackle them: new skills to develop, new tools to increase productivity, improving and mastering current processes. - book time every few months to evaluate what should be develop next, what is not being done well, what could help more, etc…

References:

These are learning from mostly on the job sources and my own learning, not all my original ideas.