NAME: weeklydata.txt

TYPE: Observational study

SIZE: 15,926 observations, 9 variables

ARTICLE TITLE: African Conflict and Climate Data for an Undergraduate Research Project

DESCRIPTIVE ABSTRACT:
These data are records of violent political conflict events on the African continent.  The records contain various descriptives of the conflicts as well as temperature information for the day from the nearest weather station.  The temperatures have also been categorized based on the behavior of temperatures
at a given location. The data have been aggregated by week and weather station.

SOURCES:
This dataset was created as a result of a merging process applied to conflictdata.txt. 

VARIABLE DESCRIPTIONS:
The data are tab-delimited.  The variable descriptors are below.

Year: year of conflict event as assigned by ACLED

Week: The week in a given year (1-53).  The R function 'strptime' was used to calculate the day of the year from the YrMoDy field in conflictdata.txt.  This value was divided by 7 and rounded up to the nearest integer to obtain a week number.

StationID: weather station identification code as assigned by the GSOD

BattlesCount: total number of conflicts categorized as Battles during the week (each event in conflictdata.txt was assigned a week number and the total events categorized as 'Battles' for the given week, year, and stationID were calculated)

VACCount: total number of conflicts categorized as Violence against civilians during the week (each event in conflictdata.txt was assigned a week number and the total events categorized as 'Violence against civilians' for the given week, year, and stationID were calculated)

RiotsCount: total number of conflicts categorized as Riots during the week (each event in conflictdata.txt was assigned a week number and the total events categorized as 'Riots' for the given week, year, and stationID were calculated)

TotalCount: total number of conflicts during the week (sum of BattlesCount, RiotsCount, and VACCount for a given year, week, and stationID)

HighMaxTemp: highest daily maximum temperature for the week assigned to the conflict event in degrees Fahrenheit (for the events with the same year, week, and stationID, this is the maximum value of the field MaxTemp in conflictdata.txt)

TempCat: the temperature category {cold(C), medium(M), or hot(H)} belonging to the HighMaxTemp value above (for given weather station (stationID), temperatures below/above one standard deviation below/above the mean for that station were classified 'cold'/'hot'; other values were classified 'medium')


STORY BEHIND THE DATA:
These data were used for an undergraduate research project to investigate possible relationships between climate data and violent political conflict.  Multiple hypotheses can be posed in a classroom or small group setting.  This is a large dataset with many possible analysis method options (regression, chi-square, anova, etc.).  These data are current and used for actual socio/economic/climatological research.  

PEDAGOGICAL NOTES:
These data can be used to teach ANOVA, linear regression, t-tests, and chi-square tests.  They invite multiple questions regarding independence of data and population definition and other statistical methodological assumptions (normality, confounding variables).  The large size of the dataset also shows that statistical significance can be found when the results are not necessarily practically significant.

SUBMITTED BY:
Name: Darcie Delzell
Affiliation: Wheaton College
Surface address: 501 College Ave. Wheaton, IL 60187
e-mail address: darcie.delzell@wheaton.edu