Results 1 to 6 of 6

Thread: [Biostatistics] Charts

  1. #1

    Default [Biostatistics] Charts

    The pie chart

    Each segment (slice) of a pie chart should be proportional to the frequency, or more helpfully the percentage of the category it represents.

    The first segments start at 12 o'clock, which is a good practice, and helps if you are comparing two or more pie charts.

    Incidentally, using the appropriate software, a good many charts can be formatted in 3D.

    Examples:

    [Biostatistics] Charts-screen-shot-2015-12-23-at-3-21-59-pm-png

    [Biostatistics] Charts-screen-shot-2015-12-23-at-3-25-26-pm-png

    Some things you should know about pie chart

    • Easy to understand
    • Can be used for either nominal or ordinal data and occasionally for discrete metric data
    • Limited by being able to display only one variable. You will therefore need a separate pie chart for each variable you want to chart
    • When comparing two or more pie charts, the area of each pie chart should be proportional to its frequency
    • Pie charts expressed in percentage frequency terms can all be drawn in the same size
    • A pie chart can lose clarity if it is used to represent more than a small number of categories (four or five or thereabouts)
    Last edited by Janis.Y.Chen; Wed 23rd December '15 at 3:29pm.
    Clinical Pharmacy Specialist - Infectious Diseases

  2. #2

    Default

    The simple bar chart

    An alternative to pie charts for nominal or ordinal data is the simple bar chart. This is a chart with frequency on the vertical axis and category on the horizontal axis. The simple bar chart is appropriate if only one variable is to be shown.

    [Biostatistics] Charts-screen-shot-2015-12-23-at-3-36-50-pm-png

    The clustered bar chart

    If you have more than one group and you want to compare them, you can use the clustered bar chart. Figure 3.5 shows a clustered bar chart. The data are percentage leisure time activity for two groups of participants: the first group with a daily total intake of calcium of <600 mg/day, and the other with >=1400 mg/day.

    [Biostatistics] Charts-screen-shot-2015-12-23-at-3-45-43-pm-png

    Some things worth noting about bar charts:

    • Fairly easy to understand
    • Can be used for either nominal or ordinal data, and occasionally for discrete metric data
    • Have the advantage over pie charts in that clustered bar chart can show several groups at once, enabling direct comparisons to be made
    • Be sure to leave gaps between the category bars. This emphasizes the categorical (or discrete) nature of the data
    • Bar charts are best expressed in percentage frequency terms; otherwise, comparisons can be difficult
    • A clustered bar chart can lose clarity if it is used to represent more than a small number of groups (five or six, or thereabouts).
    Last edited by Janis.Y.Chen; Wed 23rd December '15 at 4:12pm.
    Clinical Pharmacy Specialist - Infectious Diseases

  3. #3

    Default

    Charting discrete metric data

    We can use bar charts to graph discrete metric data in the same way as with ordinal data.

    [Biostatistics] Charts-screen-shot-2015-12-23-at-7-07-25-pm-png
    Clinical Pharmacy Specialist - Infectious Diseases

  4. #4

    Default Charting cumulative data

    The cumulative frequency curve with discrete metric data

    Cumulative frequency can be plotted for ordinal cumulative data but is more often used with metric data so that's what we'll concentrate on. The approach for discrete and continuous metric data is a little different - we will start with the discrete data case.

    The cumulative frequency chart with discrete data is known as the step chart, for which the most frequent application is with survival analysis.

    Figure 3.19 shows percentage cumulative frequency for the number of lesions among patients with a biodegradable stent. The corresponding percentage step chart is shown in Figure 3.20. Hopefully, the chart is fairly easy to understand, but basically, each time there is an increase in cumulative percentage, the step chart steps up by the amount of that increase.

    [Biostatistics] Charts-screen-shot-2015-12-23-at-7-58-04-pm-png

    The cumulative frequency curve with continuous metric data

    With continuous metric data, which is assumed to be a smooth continuum of values, you can chart cumulative frequency with a correspondingly smooth curve, known as a cumulative frequency curve or an ogive. As an example, Figure 3.21 is a cumulative frequency curve from a study of the relative efficacy of paclitaxel-eluting balloons, paclitaxel-eluting stents, and balloon angioplasty, in the management of restenosis in patients who have received a drug-eluting stent 6-8 months previously.

    The curves show the cumulative frequency of the diameter stenosis (percent) for each of the three treatments. The diameter stenosis (percent) is a measure of how much the vessel has narrowed following the insertion of the stent - the higher the percentage is, the worse will be the result. As you can see, balloon angioplasty (top curve) gave the worst result: 25 percent of these patients had a diameter stenosis of about 72 percent or more, when compared to only about 50 percent or more for those with either paclitaxel-eluting balloons or paclitaxel-eluting stents, and 75 percent of ballon angioplasty patients had a diameter stenosis of about 32% or more when compared to about 22% or more for patients with either placlitaxel-eluting balloons or paclitaxel-eluting stents.

    [Biostatistics] Charts-screen-shot-2015-12-23-at-8-14-38-pm-png
    Last edited by Janis.Y.Chen; Wed 23rd December '15 at 8:16pm.
    Clinical Pharmacy Specialist - Infectious Diseases

  5. #5
    PharmD Year 1 TomHsiung's Avatar
    Join Date
    Jan 2013
    Location
    Chengdu, Sichuan, China
    Posts
    593

    Default Charting time-based data - the time series chart

    If the data you have collected are from measurements made at regular intervals of time (minutes, weeks, etc.), you can present the data with a time series chart. Usually, these charts are used with metric data but may also be appropriate for ordinal data. Time is always plotted on the horizontal axis and data values on the vertical axis.

    As an example, Figure 3.22 shows the rate of knee replacement for men and women in the UK between 1991 and 2006. This chart is pretty much self-explanatory, but we can notice an acceleration in the rate from about 2000 onwards.

    [Biostatistics] Charts-screen-shot-2015-12-29-at-9-32-33-pm-png
    Last edited by TomHsiung; Tue 29th December '15 at 9:35pm.
    B.S. Pharm, West China School of Pharmacy, Class of 2007, Health System Pharmacist, RPh. Hematology, Infectious Disease.

  6. #6
    PharmD Year 1 TomHsiung's Avatar
    Join Date
    Jan 2013
    Location
    Chengdu, Sichuan, China
    Posts
    593

    Default Summary

    Figure 3.24 may help you to decide on the most appropriate chart for any given type of data.

    [Biostatistics] Charts-screen-shot-2015-12-29-at-9-37-24-pm-png
    B.S. Pharm, West China School of Pharmacy, Class of 2007, Health System Pharmacist, RPh. Hematology, Infectious Disease.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •