This is Tugce from Victoria University Melbourne Australia. Me and my team mate Luke did some researches on domestic violence and prepared some visualizations in SAP Lumira for analysis.
The aim of this project was gathering and exploring domestic violence data and present a story of findings. There is many sources of domestic violence results, but few sources of raw data for analysis. The majority of crime data was found to include assault statistics as opposed to raw domestic violence data.
The type of data that was collected included:
- World Bank data – excel tables
- Crime Victimisation data from the Australian Bureau of Statistics, 2008/2009 to 2013/2014.
- NSW police statistics – pdf and excel documents
- Qld police statistics – pdf format. The Queensland raw data required the installation of software in order to obtain the documents.
- SA statistics – limited pdf data
- WA – pdf data 2005 to 2012
- ACT – online, interactive infographic but no easy access to crime statistics
- Victoria – pdf police data and excel documents with domestic violence data
- Victorian post code latitude and longitude data
- Country latitude and longitude data
- LGA Country of Birth data from ABS (2011)
The Victorian provided a detailed excel document with domestic violence data by Region/PSA/LGA by year from 2009 to 2014, therefore it was decided that the focus would be on the Victoria state data. The other state data was difficult merge together and did not provide explicit domestic violence data, rather they predominantly contained assault crime data only.
Domestic Violence Around the World
The World Bank data had a completed survey showing the beliefs of women who felt it was justified for their husband/partner to beat them for a variety of reasons.
The five reasons included:
- Argues with him;
- Refuses to have ***;
- Burns the food;
- Goes out without telling him; or
- When she neglects the children.
The data included one entry per country covering the years 2005, 2006, and 2007 and provided an ideal starting point for the exploration of domestic violence. The use of these five reasons in the Australian culture would be a weak justification for domestic violence and therefore the high percentages represents a differing of cultural acceptance of domestic violence.
The three years could be viewed as a single point in time if one considers the survey as a reflection of the cultural views of the country. The cultural views of women in the country would rarely change drastically over time. That said, the findings for the few countries that were targeted by the survey were used as a guide for more in depth analysis. Only countries that showed a majority (>50%) of justification of beatings were investigated further.
Of the few countries investigated with a majority % proportion believing domestic violence was accepted, only two countries, Vietnam and Somalia, showed in other country-of-origin statistics supplied by the Victorian Police in the same time period (05/06). The country-of-origin statistics from the Victorian Police only show assault charges, not specific domestic violence related data, therefore only tentative links can be assumed between the two datasets.
The expected outcome was that there would be a reduced reporting of domestic violence incidents by people with these country of origins. As it turned out however, the Vietnam and Somalia incidents rates turned out to be above (Somalia at 22%) and below (Vietnam 10%) the assault rate average of 15% in Australia.
(The Australian average could be taken as approximately 15% assaults and this includes all countries recorded. It should be noted however that the ‘Other’ country group recorded approximately a 15% average anyway.)
The reasoning behind the figures being above and below the average may be two-fold, firstly, as stated, the cultural acceptance may lead to reduction in reporting assault incidents, especially domestic violence of husband/partners against their partners. The alternative for the elevated ratio of assaults compared to other crimes could be that there is an increased acceptance of the violent behavior which leads to an increase of these types of crimes.
The subsequent investigations into country-of-origin influence on domestic violence statistics did not yield any further insights. For example, there are a number of LGAs with higher populations of Vietnamese people households (roughly 9%), but there is no indication this has an effect on the domestic violence reporting figures, either up or down.
The two visualizations below show little correlation between country-of-origin factors and domestic violence rates. The filtered visualization shows the LGA sectors with a higher number of Vietnamese people (Greater Dandenong) is similar in domestic violence rates with the surrounding LGAs (Monash, Kingston, Knox, Casey, and Frankston).
Domestic Violence in Victoria
In Victoria, domestic violence is an issue that is growing and is a particular problem in rural areas. The visualisation below gives a summary of domestic violence incident rate changes over time and with respect to changes in population. There has been a population growth but it has been outstripped by the growth in domestic violence rates.
The areas that have the greatest rate of domestic violence can be found in the regional areas of Victoria. These LGAs include Campaspe, Central Goldfields, LaTrobe, Mildura, and Swan Hill. These areas suffer from domestic violence rates that are sometimes worse than double the average (of approximately 1000 incidents per 100,000 people population). The map data visualization shows a concentration on the regional areas.
The data had to go through a number of iterations before being made usable by the Lumira program. For instance, the incident report on domestic violence from 2009 to 2014 had to be re-organised to allow for importation. The population values were also back calculated in order to obtain accurate rate per 100,000 people. The averaging of the rate per 100k people figure would have been an inaccurate representation of the data.
Other examples of data manipulation that was required to create the set of visualization Lumira include:
- The World Bank Data cleaning and summation;
- LGA Country-of-Origin (Vic Police) extraction from PDF, formatting and Assault % calculations;
- LGA Country-of-Origin (ABS Data) post code, latitude and longitude lookup tables and summary, convert to upper case to allow merging of LGA data, percentage Vietnamese and Somalian calculations
Some other challenges faced include mapping graphical representations. On a world scale, the map works with latitude and longitude data however at a more granular view, the map is not detailed enough to give a very good visual. Below are two map visualizations, one a global scale and the other focusing on the state of Victoria.
You may find the detailed information about the project from the following video:
Mrs. Tugce Guler
SAP HANA / DW Consultant