(1) Please make and display your name cards!
(2) Make sure you're on Slack:
https://ledeprogram.slack.com/
(3) Keep PollEverywhere open
If you have any questions, just raise your hand 🖐!
Welcome! Let's get rolling!
Today
- Homework Review: Feedback / Comments
- Pair Programming: Assignment 2 / 3
- Short Lecture: Interrogating data and avoiding pitfalls
- Pitch Sharing: Let's explain our pitches to one another
Homework Review
Homework Review: Decoding Feedback
https://docs.google.com/document/d/1VIWR76f-EPa_3gOUv2F1DlIZvpONntoLyxK8cOwfKd0/edit
Homework Review: Decoding Feedback
https://docs.google.com/document/d/1VIWR76f-EPa_3gOUv2F1DlIZvpONntoLyxK8cOwfKd0/edit
Please note that “⛔️” denotes errors that require a response from you.
Your response can be a rewrite of that section or a reflection or discussion in the comment to demonstrate that you understand thoroughly what the issue was and how to prevent it from happening again.
How're things going in California?
Live Election Results
https://www.washingtonpost.com/elections/election-results/california/governor-recall/
Pre-election forecast
How're things going in...Germany?
Election: Sunday 9/26
- Background Info. What is this election about?
- How do German elections work? Ground rules.
- How do we have to think about polling data differently in a parliamentary system?
Survey
How is Assignment 2 going?
Interrogating Data And Avoiding Pitfalls
Descriptive Statistics
(recap)
Summary Statistics
Measures of Central Tendency
- Mean = μ
- Median
- etc...
Measures of Variability
- Range
- Standard Deviation = σ
- Variance = σ2
- etc...
Do
- "slice and dice" the data, calculating summary statistics of various subsets (pivot tables are a great way to do this)
Don't
- Only calculate summary stats on overall data given to you
Summary Statistics
Do
- Be aware of what you're losing when you calculate a summary statistic
Don't
- Rely on the summary statistic alone without considering the underlying data
-
are there outliers?
-
how is the data distributed?
-
how many data points are there?
-
etc...
-
- Assume your data is normally distributed
Summary Statistics
Do
- Pay attention to time
Don't
- Just average all the data
Summary Statistics
Do
- Look at particular datapoints in the context of all the datapoints and provide that context to readers ("This 150 gram apple is the heaviest one in the barrel, by far")
Don't
- Look at datapoints alone ("This apple weighs 150 grams")
Summary Statistics
Plotting
As a form of data exploration
Plotting
As a form of data exploration
- Scatter Plots - to see the data points
- Scatter plots or line charts to see trends over time
- Scatter plots to see correlations between two columns
- Bar charts to compare discrete values
- Histograms to look at distributions
- Line charts to see timeseries data
- and more...
Tools:
- Excel or Google Spreadsheet Charts
Do
- Plot
Don't
- Rely on the plot alone
Plotting
As a form of data exploration
Do
- Plot subsets of the data
- Make more than one plot
Don't
- Only plot all the data in one chart type
Plotting
As a form of data exploration
Pay Attention To Distribution Of Data
Other Tips
- Start with general exploration but quickly move on to asking targeted journalistic questions of the data
- "Are state polls less accurate than national polls?"
- "Which states have the lowest income?"
- "Why is Chicago an outlier?"
-
Make sure to interrogate your editorial decisions
- What variables have I pivoted by?
- Should I be pivoting by something else?
- Should I be filtering the data somehow?
- What information do I lose by summarizing the data this way?
- Am I considering the temporal element in the data?
- What variables have I pivoted by?
- Keep in mind what the data set cannot tell you. What are the limitations of the data and of your approach?
Quantitative Editing
- Data Diary
- How exactly did I get from the source data to what I have presented? Is my work reproducible by an editor?
- Raw Materials
- Documentation
- Instructions
- Make methodological assumptions explicit
- How exactly did I get from the source data to what I have presented? Is my work reproducible by an editor?
- Video Talk: What is quantitative editing?
- Laura Bronner's guide to submit an article for quant edits
Pair Programming
Pair Programming
Pair Programming
Driver
Has the computer/keyboard. Is writing the formulas, charting data, etc etc....
Navigator
Thinking ahead, looking up how to do something, listening to the driver and bouncing around ideas...
Pair Programming
Driver
Has the computer/keyboard. Is writing the formulas, charting data, etc etc....
Navigator
Thinking ahead, looking up how to do something, listening to the driver and bouncing around ideas...
Pair Programming
Switch it up occasionally
Pair Programming
Pair Programming
Rubber Duck Debugging
Assignment 3?
Project Preferences
Now that you've heard everyone's pitches...tell me who you are interested in working with and why
What's coming up?
- Submit your project preferences by tomorrow
- Project groups will be assigned over the weekend
- I will assign some reading, and maybe a small data project, but nothing extensive...