Democratizing Data

Charles Guthrie, data visualization engineer, talk about making The New York Times more data-driven by expanding access to internal data through simple visualization tools.

(Presented at the Data Visualization Innovation Summit)

Transcript

1. Democratizing Data Democratizing Data Making data accessible at the New York Times
2. Making Data Useful Making Data Useful Answering Questions
3. Analyst Stakeholder Database
4. How are we doing? From: Advertising To: Analysts select date, revenue from revenue_table where date>2015-01-01 and date<today What are people reading? From: Newsroom To: Analysts select article, current_page_views from streaming_event_tracker group by article
5. (Fake Data) Re: How are we doing? From: Analysts To: Advertising
6. Numbers need Numbers need context context
7. 328 homicides 328 homicides in in NYC in 2014 NYC in 2014
8. source: http://en.wikipedia.org/wiki/Crime_in_New_York_City#cite_note-one-177
9. (Fake Data) Re: How are we doing? From: Advertising To: Analysts
11. "Email Guilt" "Email Guilt"
12. Meeting Agenda Meeting Agenda "Get on same page" Read every bullet on slides out loud Review stuff we already know Tangential conversation with one person
13. Push Push
14. Pull Pull Easy Answer Machine
15. Building the Building the Easy Answer Machine
16. Requirements Requirements Easy to use Visual Answer questions Up-to-date
18. Identify and Identify and answer the main answer the main question first question first
19. Who's winning the election? Who's winning the election? http://elections.nytimes.com/2014/results/senate?utm_source=top_nav&utm_medium=web&utm_campaign=election-2014
20. Re: What are people reading? From: Newsroom To: Analysts tutorial at: http://bl.ocks.org/charlesdguthrie/11356441
21. How are we doing? From: Advertising To: Analysts Do I need to be worried? How are we doing?
22. Re: How are we doing? From: Advertising To: Analysts
23. Matching the chart Matching the chart to the question to the question
25. relationships between two variables trends and changes; historical context simple comparisons comparisons with historical context Multi-Line Chart Multi-Line Chart
26. Comparing values with historical context source: http://www.nytimes.com/interactive/2009/11/06/business/economy/unemployment-lines.html
27. (Fake Data) Comparing values with historical context
28. Answering Answering questions questions without without charts... charts...
29. Re: How are we doing? From: Advertising To: Analysts
30. Re: How are we doing? From: Advertising To: Analysts (Fake Data)
31. Answering Answering questions questions without without charts... charts... ...or with nothing at all ...or with nothing at all
32. Nothing to tell Nothing to tell Do I need to be worried? From: Advertising To: Analysts
33. Alert! Revenue dropped by 27% From: Dash To: Advertising
34. "OK, why?" "OK, why?" Follow-up Questions
35. Overview first, Overview first, zoom and filter, zoom and filter, then details-on-demand then details-on-demand - Ben Shneiderman - Ben Shneiderman
36. Election: How close is it? Election: How close is it? http://elections.nytimes.com/2014/results/senate?utm_source=top_nav&utm_medium=web&utm_campaign=election-2014
37. Re: What are people reading? From: Newsroom To: Analysts How do people reach our articles?
38. (Fake Data)
39. WHY HAS REVENUE DROPPED BY 25%?!? From: Advertising To: Analysts (Fake Data) Re: How are we doing
40. Why is revenue down? Why is revenue down? Zoom Filter Filter Details on Demand
41. Demo Demo http://localhost:3274/#/site-activities
42. Data, Democratized Data, Democratized Results
43. Data, Democratized Data, Democratized Page Views Visitors Homepage stats Products Video Subscriptions Crosswords Advertising Personalization A/B Tests Data Governance URLs Opinion Top Articles Etc. 18 Teams Over 100 charts Crosswords
45. Freed up for more advanced problems Building out more tools Use data rather than intuition More insightful questions Analysts Stakeholders
46. Meeting Agenda Meeting Agenda (now that everyone is informed) discuss analytics insights make informed decisions
47. Things we Learned Things we Learned Answer the main question first Pick the right chart - or no chart Zoom and filter for followup questions
48. Data Awareness Data Awareness Data-Driven Decisions Data-Driven Decisions
49. charlieguthrie.net @charlieguthrie linkedin.com/in/guthriecharlie Thank you Thank you