Presentation is loading. Please wait.

Presentation is loading. Please wait.

Presenting Data Public Health Intelligence training course day 5 Liz Rolfe Knowledge & Intelligence Team (South West)

Similar presentations


Presentation on theme: "Presenting Data Public Health Intelligence training course day 5 Liz Rolfe Knowledge & Intelligence Team (South West)"— Presentation transcript:

1 Presenting Data Public Health Intelligence training course day 5 Liz Rolfe Knowledge & Intelligence Team (South West)

2 Learning Objectives At the end of the course, participants will be able to... Some basics of visualisation Identify what chart is required Key things to include Key things to avoid 2Presenting Data

3 “Excellence in statistical graphics consists of complex ideas communicated with clarity, precision and efficiency” (Tufte, 1983) Show the data accurately & clearly Avoid distortion Encourage the eye to compare different pieces of data 3Presenting Data

4 What you see is what you get? 4Presenting Data

5 Visualisation Data will not speak for itself A picture is worth a thousand words (or numbers?) 5Presenting Data

6 Visualisation & good presentation Keep it simple Do not present unnecessary information  ‘Chart junk’  Decimal places on axis labels  Repeated information All graphs must have  Titles  Labelled axis  Timeframe  Source 6Presenting Data

7 Examples 7Presenting Data

8 Bar chart – what’s wrong? 8Presenting Data

9 Is this easier to interpret? 9Presenting Data

10 Where does Wandsworth rank? How much worse than Westminster is it? 10Presenting Data Source: Office for National Statistics and Teenage Pregnancy Unit

11 Wandsworth is fifth of the 14 Inner London Boroughs; it’s rate is double that of Westminster 11Presenting Data Source: Office for National Statistics and Teenage Pregnancy Unit

12 Which is better – vertical or horizontal? 12Presenting Data Source: Office for National Statistics and Teenage Pregnancy Unit

13 Where the y axis does not start at 0 13Presenting Data

14 Presenting charts in juxtaposition: which is more powerful? 14Presenting Data Breast cancer incidence and mortality by deprivation South West 2001-5

15 Or this? 15Presenting Data Breast cancer incidence and mortality by deprivation, South West 2001-5

16 Highlight the areas of interest 16Presenting Data Directly age standardised rate of mortality due to suicide and undetermined injury among young people aged 15-24 years by LAD, 1997-2004

17 Highlight information on charts 17Presenting Data *Data are for 2005 except; Ireland, England, Wales, New Zealand – 2007, Austria, Finland, France, Italy, Mexico, Netherlands, Spain – 2006, Belgium, Poland, USA – 2004, Greece, Hungary, Turkey, Scotland – 2003, Iceland, Switzerland -2002, Australia, Portugal -1999, Obese defined as BMI ≥ 30kg/m2. http://www.noo.org.uk/NOO_about_obesity/international/

18 Clustered and stacked bar charts are used to describe two or more categorical variables simultaneously. 18Presenting Data Group UnclassifiedNormalIGTNew Diabetes Percent 100 80 60 40 20 0 Gender Male Female Group UnclassifiedNormalIGTNew Diabetes Count 300 200 100 0 Gender Female Male Clustered bar chart showing the percentage in each group within each gender. Stacked bar chart showing the number of people of each gender, within each of four groups.

19 100% stacked vertical bar chart 19Presenting Data Difficult to make comparisons among the second, third, or subsequent segments in a stacked bar because judgment is not being made from a common base. Which region has largest proportion of males 35- 64?

20 Scarf chart 20Presenting Data The scarf chart is a special case of the stacked vertical bar. Comparisons are easier, but, where possible, avoid too many comparisons

21 3D pie chart 21Presenting Data You can try various things to spice up the chart. The inclination and distortion that produce the 3D effects make it difficult to interpret the data. In the 3D view, the Hazelnuts (4%) and the Cashews (8%) look similar.

22 3D bar chart 22Presenting Data What % of abortions at 13+ gestation in East Midlands?

23 23Presenting Data What % of abortions at 13+ gestation in East Midlands? What % of abortions in England at gestation< 10 weeks?

24 24Presenting Data Adapted from age-class prevalence rates for selected conditions from the Health Survey for England, 1991-97. Self reported "very bad health” Asthana, S et al. J Epidemiol Community Health 2004;58:303-307Copyright ©2004 BMJ Publishing Group Ltd. men women Is this 3-D chart informative?

25 Does adding pictures make the graph more powerful? 25Presenting Data

26 Lines or bars? 26Presenting Data Length of stay as an inpatient 3728211915131197531 Count 40 30 20 10 0

27 What’s wrong here? 27Presenting Data

28 Line chart – what’s wrong? 28Presenting Data Source: Office for National Statistics and Teenage Pregnancy Unit

29 Is this easier to interpret? 29Presenting Data Source: Office for National Statistics and Teenage Pregnancy Unit

30 Line chart 30Presenting Data Life Expectancy at Birth Target by 2010 78.6 years for men 82.5 years for women LE Gap = 5 years Useful for looking at trends of time

31 Scatter charts 31Presenting Data What information does this show?

32 Presenting confidence intervals 32Presenting Data Breast cancer incidence by deprivation, South West 2001-5

33 Further ways to present confidence intervals 33Presenting Data

34 Further ways to present confidence intervals 34Presenting Data

35 35Presenting Data

36 Drawing Good Graphs If the “story” is simple, keep it simple. If the “story” is complex, make it look simple. Tell the truth – don’t distort the data. 36Presenting Data

37 Further examples 37Presenting Data

38 Spidergram can compare several variables simultaneously 38Presenting Data

39 Boxplots are graphical representations which clearly show the quartiles and extremes or outliers of a set of data in a compact manner. 39Presenting Data

40 A form of boxplots are used in Health Profiles 40Presenting Data

41 Quadrant analysis 41Presenting Data

42 Dashboards 42Presenting Data

43 Data in tables 43Presenting Data Network1999/002000/012001/022002/032003/042004/052005/06total Anglia276929053319328433893938387623480 B&H143717762149197520482296250314184 East of England565266317207 73178516927151801 Essex144619501739194818802282289214137 Network1999/002000/012001/022002/032003/042004/052005/06 Total Anglia2,7692,9053,3193,2843,3893,9383,87623,480 Beds & Herts1,4371,7762,1491,9752,0482,2962,50314,184 Essex1,4461,9501,7391,9481,8802,2822,89214,137 East of England5,6526,6317,207 7,3178,5169,271 51,801 Not so clear Clearer? Total number of revascularisation procedures (PCI, coronary angiography and CABG) by cardiac network in the East of England: Source: HES online data

44 Learning Objectives Some basics of visualisation Identify what chart is required Key things to include Key things to avoid 44Presenting Data

45 References Dibben, C, Watson, J., Smith, T., Cox, M., Manley, D., Perry, I., Rolfe, L., Barnes, H., Wilkinson, K., Linn, J., Liu, L., Sims, A., and Hill, A. (2008) The Health Poverty Index. The NHS Information Centre. Leeds, UK Wainer, H., How to display data badly. The American Statistician, May 1984, Vol 38, No 2. Tufte, E., (1983) The visual display of quantitative information, Graphics Press, Connecticut, USA http://www.neighbourhood.statistics.gov.uk/dissemination/Info.do?page=use rguide/detailedguidance/statisticalpresentation/best-practice-key- principles.htmhttp://www.neighbourhood.statistics.gov.uk/dissemination/Info.do?page=use rguide/detailedguidance/statisticalpresentation/best-practice-key- principles.htm http://www.apho.org.uk/default.aspx?QN=P_HEALTH_PROFILES http://www.noo.org.uk/uploads/doc168_2_noo_NCMPreport1_110509.pdf tthttp://www.noo.org.uk/uploads/doc168_2_noo_NCMPreport1_110509.pdf http://www.gapminder.org/ 45Presenting Data


Download ppt "Presenting Data Public Health Intelligence training course day 5 Liz Rolfe Knowledge & Intelligence Team (South West)"

Similar presentations


Ads by Google