Introduction to Protobi

Here's a brief tour of Protobi™ using public health data from the CDC National Ambulatory Medical Care Survey (NAMCS), an annual survey of US outpatient MDs who complete detailed chart reviews for patients they see.

Scenario toolbar

The Scenario Toolbar shows the current sample size and how data are displayed. Hover over the buttons for more information.

N= shows the current sample (i.e. {{getBasis()}}). Click to see current filters.
Clear resets current filters returning to all records.
Baseline sets the current sample as the baseline for comparisons.
# | % toggles whether to show frequencies as counts (#) or percents (%).
Formats toggles display of value labels vs raw data.
Missing toggles whether missing (blank or NA) values count toward percents.

Interactive elements

Protobi ™ displays data as interactive charts:

Click a value to drill in.
Click again to drill out.
Shift+Click to select NOT a value.
Option+Click to select OR a value.

Constrained values are highlighted in gold.

See the N size in the Scenario toolbar update. All elements update as you set constraints.

The baseline distribution is shown as a light grey shadow for reference. Arrow icons highlight significant differences from baseline.

General/family practice Internal medicine Pediatrics General surgery Obstetrics and gynecology Orthopedic surgery Cardiovascular diseases Dermatology Urology Psychiatry Neurology Ophthamology Otalaryngology Other specialties Oncology

Numeric variables

Numeric values are automatically rounded, with means presented. Categorical variables can have value labels for readability.

Change the rounding bin size by selecting the gear icon.

Positive number (e.g. 15)
auto to autoround
0 or blank to turn rounding off

Means are shown for numeric variables.

Statistical comparisons

The baseline distribution is shown as light grey shadows for reference. You can set the baseline at any point by clicking the "Baseline" button in the Scenario toolbar.

Triangle icons (i.e. ) highlight statistically significant differences from baseline (using a Z-, F-, or chi-square test as appropriate)

For instance, click on 'Pediatrics' above and observe that most patients visiting a Pediatrian are under age 20. And more likely to be male.

Interactive crosstabs

We've taken a distinctive approach to the traditional crosstab, making it instantly update as you explore, and allow you to click within it to set new findings.

The way you style an element (i.e. formats, value labels, rounding) determines how it appears as a crosstab. So you're not "programming crosstabs", you're exploring the data and getting crosstabs that reflect it.

Statistical significance tests compare each column against the average of all other columns.

Value labels and missing values

Toggle between displaying formatted and raw values using the Format toolbar button (shortcut 'f').

Protobi also provides nuanced handling of missing / blank / skipped values.

Toggle missing values from percentages using the Missing toolbar button (shortcut 'm').

Embeddable JS Library

For developers

Protobi is a lighweight JS library you can embed and remix in your data portal.

Notice that all of the above charts can be specified in HTML and/or Javascript enabling you to create your own designs. In our app or yours.

Male Female White only Black/African American only Asian only Native American/Oth Pac Isl only American Indian/Alaska Native only More than one race reported Blank General/family practice Internal medicine Pediatrics General surgery Obstetrics and gynecology Orthopedic surgery Cardiovascular diseases Dermatology Urology Psychiatry Neurology Ophthamology Otalaryngology Other specialties Oncology