Presentation is loading. Please wait.

Presentation is loading. Please wait.

Welcome to Periscope Data

Similar presentations


Presentation on theme: "Welcome to Periscope Data"— Presentation transcript:

1 Welcome to Periscope Data
Data Teams, Unite! Sean Cook, Community & Content Nov 2, 2017

2 We Love Wall-Mounted Dashboards
Great to track KPIs, user commentary and sales metrics. A fun place to try new design ideas. A great use of Periscope Data Shared dashboards! We have a blog post!

3 Rasberry Pis get it Done
Cheap, easy and fun to setup! For 35 dollars a screen you can have a full computer and all that entails With Linux images you only need to build once Our cases were made by C4 Labs for ~$25 Use a good power supply. Buy the newest model. Use NOOBS & Chromium Can be set to boot, load a dash in full screen and hide the cursor in one set of commands!

4 Use Python too For More Fun
Schedule rotation of dashboards Refresh dashboards Change dashboards from your desk Display Dashboards based on criteria

5 Use Firebase to Change Dashboards from your Desk
Google Firebase is a cloud database for applications with a free tier that more than covers any url storing needs. Use Python to check periodically Change dashboards from your desk or home Or put on the World Cup/World Series/Netflix!

6 Chromium Extensions User CSS Change CSS for any site on the fly.
Have fun, spruce up your dashboards, or be festive! Revolver Awesome way to smoothly rotate tabs!

7 https://app. periscopedata

8 https://app. periscopedata

9 User CSS can also modify the Periscope Editor

10 SQL For Visual Analysis Britton Stamper - Periscope Solutions Engineer

11 Visual Analysis Q: What is a chart?
A: Charts can explore phenomena in data and make data easier to consume. Design Goals: Identify the important trends quickly and easily Open up new types of analysis Nudge the user to proper analysis

12 Which is more valuable to the consumer?

13 Which is more valuable to the consumer?

14 Which is more valuable to the consumer?

15 The Components of an Advanced Chart
The most important consideration is the alignment between use-case and design To create any valuable chart, the analyst must take into account: Properly structure the data set Chose the series type by common concepts Color use as a tool Feedback from the consumer Series = Positive/Negative Data = value / lag(value) over(order by date) - 1

16 General Method for Designing Data Sets
Define the goal behind the visual (comparison, relationship, distribution, etc.) Choose the types of values need to be returned (ex. static/dynamic values over time) Define the different series necessary using labels What structure will the data need to achieve these?

17 Exercise - DIY Histogram
Show distribution Count by bin We can use mod truncate to the minimum value per bin to generate the dataset What is the goal? How can we structure dataset? What transformations are necessary?

18 Exercise - DIY Histogram
select value - mod(value,5) as value , count(1) from data group by 1

19 Parameterized Snippets (Don’t Repeat Yourself)
Periscope offers Parameterized Snippets that can do all the heavy lifting for you Once you have a great new chart type, you can generalize the transformations and create simplified references to repeat them over and over again A parameterized snippet takes the form:

20 Parameterized Snippets - Examples
Each of these snippets will transform the common results of a table into a data set ready for a specific chart type:

21 Everything to Date

22 Waterfall Chart Shows a running total over a series of values. Allows a consumer to quickly see the rise and fall over the series so they can identify the points in the chain causing a large drop. This can also be a great way to display a KPI across another dimension. Base: Bar Chart For the two running total series, color = #f9f9f9 Set the negative series to red, a possible color = #fb3a2d Parameterized Snippet Name: waterfall_charts(table,value,order) select [order] , abs([value]) , case when [value] > 0 then sum([value]) over(order by ord rows between unbounded preceding and 1 preceding) else [value] + sum([value]) over(order by ord rows between unbounded preceding and 1 preceding) end as running_total when value > 0 then 'positive' else 'negative' end as sign from [table]

23 KPI vs Benchmark over Time
Shows a value over time versus a fixed benchmark. If the benchmark changes, a union can be used to break up multiple select statements that return the unique bounds for each benchmark change. Base: Bar Chart Set the Benchmark series to Line select kpi , 9 as benchmark , date from table

24 It’s on the Community!

25 Thank you!


Download ppt "Welcome to Periscope Data"

Similar presentations


Ads by Google