GEOG 3MB3 Lecture Notes - Lecture 3: Zip Code, Arson
28 May 2018
School
Department
Course
Professor
September 16, 2016
Manipulating Data
- In the real world you run into situations where you have data but you must make
it analytically useful
- Data tables are convenient but often our data is not in this form or it is just not
right
- Often you are working with data and there is a table that you can’t even see (I
don’t see how this was relevant)
Getting data into the right format
- There is a monitoring station in this provincial park that takes samples every half
hour.
- This may be too much data to be useful, so they may only look at one
measurement at a given time of day
Aggregation
- If you don’t want the data on a fine level you can instead just look at the data per
day
- For instance, these 3 date time/points on the left will aggregate to 1 data day/
point
- In this example taking the average/mean was used to get the 1 data point from 3
- They could have used other functions, they could have counted them, the many
is being reduced to a single thing
- You could just say “I am grouping” instead of “taking an aggregate”
Agg example
- These look at arson cases in the US
- He aggregated based on zip code in order to look at where the most arsons are
occurring
Agg example 2
- This data is looking at sediment in water samples
- He may say I don’t need this many measurements why don’t I just take an
average instead?
Getting data into the right format – Aggregation
find more resources at oneclass.com
find more resources at oneclass.com
Document Summary
In the real world you run into situations where you have data but you must make it analytically useful. Data tables are convenient but often our data is not in this form or it is just not right. Often you are working with data and there is a table that you can"t even see (i don"t see how this was relevant) There is a monitoring station in this provincial park that takes samples every half hour. This may be too much data to be useful, so they may only look at one measurement at a given time of day. If you don"t want the data on a ne level you can instead just look at the data per day. For instance, these 3 date time/points on the left will aggregate to 1 data day/ point. In this example taking the average/mean was used to get the 1 data point from 3.
