Spatial Mapping of Probability Distribution

This view contains the floor plan of the Mock-up showing each rotatable element at the angles previously selected in the radial histograms.
The opacity for each drawn element reflects the probability of it being in that state. The probability depends on the Scale that is set: frequency or duration of a state.
The probability p of an element j at angle i with value α_i (frequency or duration at angle i) is calculated as follows:
$$ p_j(\alpha_i) = 1-c_j^{\frac{\alpha_i}{max_j}} \quad, \text{ and } max_j = max \{ \alpha_i | \text{ angle $ i $ is selected for element $ j $\}}$$
Filter filter_list

Clicking on the grey bar with the title Filter reveals the filters that can be set. To apply them onto the data, click on the symbol.

Double Grouped Radial Histogram

This view contains a grouped histogram for each rotatable element with a radial x-axis representing the orientation of the element.
It consists of two parts: an outer and inner radial histogram. The outer is fixed and only changes when new filters are applied. It shows the distribution of frequency and duration of states of the element over all angles. The inner changes through selection of the bars of all four histograms, and its bars are coloured darker. It also shows the distribution of frequency and duration of states of the elements but only over selected angles.
There are three modes that can be set within the filter:

• Angle grouping size: Set the bucket size in degrees by entering an integer between 1 and 90 and pressing ENTER. The default is 8 degrees.
• Scale: Choose either Frequency of States or Duration of States for the radius.

• Hovering: When hovering over a bar, that bar is highlighted in orange and the exact Scale value and angles are shown below the radial histogram.
• Selection: A single bar can be selected by clicking on it. Multiple bars can be selected by dragging over bars while holding the left mouse down.
  By selecting bars one restricts the states of the respective element that are considered for the inner radial histograms. That is, in all inner radial histograms only frequency and duration of the set of states of the Mock-up are displayed, where the elements are in either of the selected states (if there are any).
  Generally, the more bars are selected the slower the process becomes.
• Deselection: This is done by simply clicking on the bars again. If all bars are to be deselcted for a histogram, use the highlight_off button instead.

Note, each histogram is coloured differently. The same colour is used to draw the elements in the spatial mapping. Also, any angle measured by the sensors that is outside the pre-defined range of the element was truncated to 0 or to the max possible angle.

View per Weekday

This view shows the Double Grouped Radial Histograms for each weekday and each rotatable element. All bars are linked together for the histograms of the weekdays and the main historgram above.

While the radial histograms show how often/long different states of the Mockup could be observed, it does not show in what weeks.
This view contains a bar chart that shows just that. A bar shows the percentage of states (either in frequency or duration) of the total observed states that were observed in the given week. The total observed states are defined through selection in the radial histogram. There is one bar chart for each group. When hovering over a bar, it shows the exact Scale value and percentage.
Note, this chart is empty if no bars are selected.

Every time the modes are changed or the filters on top are newly applied, the symbol must be clicked in order to load/updated this view.
Mind, this visualization, especially the inital loading, can take a few seconds to load.

Most Common Words
This view shows the ranking of the top 15 most common words. For this, every word of a given wordlength is extracted from the string and compared with each other.
There are various modes that can be set:

• wordlength: The length of a word, i.e. the length of a sequence of actions by a test-subject.
• maximal distance: The maximal allowed distance of a word to count as an occurence. The distance is computed using the Levensthein Edit distance.
• letter per: Every sensor either maps uniquely to a letter (sensor) or every sensor in the same room map to the same letter (room).
• modus: This sets how to create the long string. The normal mode applies no modification, while without-repeats removes any consecutive duplicate sensors (e.g. AABCCCD turns to ABCD). And latly no-closed removes any sensor meauserments for 'CLOSED' events.
• Time-Interval: This implementation is still slow, and in order to compensate for this, the time filters on top are NOT applied for this view. Instead, intervals of 3 hours can be selected to filter the data for.

The bars in the charts are linked to two other visualizaions, explained in the next two sections.
Below you find the list of sensor to letter mapping. The room letters have the following meaning:
R - Rotatable Element, K - Kitchen, S - Bedroom, W - Livingroom, B-Bathroom, H - Entrance.

Trace Path Approximation

After clicking on a bar in the Most Common Words chart, an approximation of the path taken for executing that word will be show in in the floor plan on the right.
Recall a word is a sequence of actions on the elements the letters corresond to.

Timeline

On click on a bar in the most common bar charts, the timeline information for the corresponding word will be displayed. It shows for every occurence counted the time a particular occurence was observed, either by week or by weekday. The x-axis shows the time of day.
Note this is but an approximation and solely based on the sensor data provided. There is on guarantee that it is the actual path taken.
Next to the y-label, the exact number of occurences is displayed in brackets.
This view is empty if no bar was selected.