What is Signals

Signals is an instrument to observe the variation of the data differentiating it from the statistical noise. Morphs are mutable objects, which can be sent to interested recipients, which encapsulate the data, the related context, the variation charts, automatic and user annotations that act as translators for the different actors interested in knowing the meaning of the march. of its critical processes.

In addition to the typical statistical signals (dominant, moderate, weak), it has other derivatives, "fuzzy" versions (based on fuzzy logic) and fractals that allow understanding at other (subjective) levels.

You can explore an entire extensible library of critical process indicators using predefined and intelligent categories using a top-level domain language, filter and annotate the data.

User Interface - Signals

With this tool you can explore the repository of measurements, manipulate, and query measurement indicators, as well as compose dashboards. The window is divided into four panels:

List of categories and indicators
Control Panel (filters)
Graphics Panel
Data Table Pane

In addition to these panels, there is a field of inquiry / edit the description of the indicator (on the control panel, light blue). There you can change the description and save changes.

Using the menu you can access all the features of the tool (many functions are also available from context menus displayed by right mouse button)

This tool can be displayed full-screen or normal window. You can enter full screen to have selected an indicator data by clicking on the corresponding icon in the system tray (top right)Fig. 26. Signals

Fig. 27. Signals full screen

Control Panel

Through this screen area you select what specific KPI & project you want to see. The panel consists of the selection lists and a project indicator, and a series of controls that act on this selection:

View project indicators, all or selected project: shows indicators in the list of those indicators as a favorite project of all or selected project respectively. You can add and remove indicators from the list of favorites using the context menu (right click) menu or the main menu bar.

Grouping indicators into categories or process areas.

See phases, trends, quarterly/weekly organizational and baseline detail: controls what is displayed in the data panel, iterations or phases of the project (in chronological order) quarterly periods (cumulative or average as measured or ratio, to see trends), averages quarterly for the same type of projects (organizational) and weekly (baseline consolidated by project type) or weekly detail (within phases)

Show only major phases: phase allows you to filter less or individuals, focusing on the typical phases. It uses a naming convention ("xyz") to detect, failing those lasting more than a month

Filter projects by type and volume of activity ("performance"). Allows you to use a cluster of qualitative projects (little, medium, long, etc..)

It may be that for certain projects or indicators has no data yet. Both the list of project indicators like you to determine where to look, with different emphasis in italics detect items with less data (or no data). This allows you to quickly select those parameters that return meaningful data

Charts and data panel

It is the area where you can see the process control charts and data table for the selected indicator through the Control Panel. In the top panel, the graphs X and mR (Normal Distribution and Variation) will show the distribution and variation of the data in graphical form for a better appreciation of patterns, while the bottom, the numerical data table, complete its view with the corresponding numerical values. Explore this table using a contextual menu (right click) menu or View menu bar to search for text data, display the minimum, maximum, or average. Meanwhile the graph X can be filtered with the "View" menu by selecting points on the graph with the mouse to mark:

Start / End series
Exclude / Include points

With this filter you can analyze sub-series, excluding special items so as not to influence the analysis.

Other ways to filter by dates are from the context menu of the data table by selecting a date / period or from the slider below the graph (drag the bar button left and right to set the date from or until and apply the filter ). This bar has a context menu in the drag button that allows to filter by from or to date, as well as remove filters. An additional option allows you to add the selected date to the global context (event dictionary).

Filters can be applied on more than one morph, you simply have to enable "synchronization" and all morphs on the desktop will apply the filter when one of them changes it. To activate / deactivate the synchronization use the option of the context menu of the desktop, or the one of an individual morph of the desktop to apply it only to said morph.

When you select to see a trend, you can "Extrapolate" to the future observed a linear or exponential extrapolation as appropriate. You can evaluate the accuracy of the extrapolation made by consulting the R2 value was obtained (considered good for values greater than or equal to 0.75). Options for extrapolating include maintaining the current mean, or modifying it + - a percentage, or simulating a Factorial Brownian motion according to the hurst coefficient (last third of data).

The graphic panel can open it in a new window every time you want, for example, to continue with another without losing the current analysis. If you want to compare several series, all on one graph, you can use the "Add extra series" option and you can compare the curves (eg multiple projects, project types, correlated indicators, etc..)

The graphical display type which has adapted to the category or type of the selected indicator. Thus, you may also see grouped data as pie charts both bar type. In these cases you can still perform a statistical analysis of variance on each cluster in particular (or total), or view the most recent period (last), the average or the duration of a "global maximum".

The data panel with graphics and data can be opened independently on the desktop, such as Morph. The desktop morphs have a "halo" menu with general options, one of which allows you to display a specific menu according to the type of panel. These morphs can be duplicated, copied to the pasteboard and copied onto a page of objects on the board. You can update all the morphs of the desktop through the menu of the desktop.

You can share the charts and exporting data table to external files or send them by email.

Fig. 28. Example of a Panel Data type indicator Bar

Charts

Charts are one of the three components of the statistical analysis system (along with data values and context) that allow variations to be visualized to help control processes by establishing their causes. There are three types of charts (lines, bars, and circles). The former are used in process control. The labels correspond to signs of variation (various types, identified with a color). It is convenient to filter the period using sliding the horizontal bar to dynamically visualize the changes. The magnitude of the variations is plotted on the curve below, called the mR graph.

With Author role: Graphics can be labeled to indicate patterns of various types. User annotations can be global or dot level, and graphic elements (text, ellipse, arrow). You can perform operations with the graph, such as adding series (extras and composing a polymetric), extrapolating, simulating, correlating. With the control panel you can change the KPI, project, view. You can add one or more panels to the Stack, or open it on the Desktop as Morphs, share it by email or via Chat with other connected users

With role Admin: From the Library, you can edit the KPI corresponding to the graph (in a Morph, visualize the object in the library with the halo option). You can obtain other types of graphs (Morphs) not related to the panel (lines, regression, limits, log, bars) from the data panel menu, or for a particular column (bars). Another way is through scripting and bundles predefined or created by you

Con role Developer: Puede inspeccionar cada morph y el modelo con las opciones correspondientes del halo

Data

The data should always be considered in conjunction with the other components of the analysis: graph and context. Numerical data (which can be aggregated) always correspond to a project (or set of projects of the same type), and a period. The data panel is displayed decorated with the colors and names of the statistical signals

With Author role: you can open the panel on the desktop as Morph to study it in detail (filter, sort, group, encode scripts in IHDSL query, open additional graphics). You can copy and export the data. It should normally be displayed next to the chart, but you can control its display or not

With Admin role: open the data panel in a new panel on the desktop and create new scripts / bundles, data source

With Developer role: you can edit the Morph of the open data panel on the Desktop via the halo and explore the data, view it and / or debug your new IHDSL script - Build your scripts using predefined webservices (see service bundles) or create new services based on the previous
Default data format (KPI)

KPIs must contain the following default fields:

    #project: name of the project (or group of projects)
    #releaseName: version name, phase, Sprint, period
    #releaseDate: version end date, phase, sprint, or period (start)

Then there must be a Value field with the numeric value of the indicator. In case the type of representative graph of the indicator is bars or circular, before comes a series of fields for each grouping (in case of circular type, exactly five groups)

Those indicators that also provide metadata, the following fields are added, related to the components of an Activity System:

    #instrumentos: type of issue
    #operations: states of the process flow
    #actors: user / user group names

Fig. 29. Example data showing patterns

Metadata

The metadata of an indicator (category "Activity System") consists of information added to its value related to the Activity System involved, and is represented by values related to its components, specifically the instruments, operations, and actors involved in the activity measured by the indicator.

The instruments in Jira are represented by the types of issues, the operations with the states of the workflows, and the actors with the users and groups of users related to the measured tickets. These components can be used as filters in the exploration of signals in the data, and in the subsequent analysis using Activity Systems.

In this way, the metadata provides additional information to the context of the indicator ready to be analyzed through the conceptual framework to determine the possible patterns of activity of a Story.

To indicate that your indicator supports metadata, remember to categorize it as "Activity System" (see examples) and qualify the RQL code with metaCode: (if you do not provide a specific code, the application will try to derive it from code: common)

Context

The context that must accompany any set of data and graphics. The context provides situational information that allows explaining and understanding the other components as a whole. It has several levels: KPI context (objective, formula), data (quality index, entropy), project (typifications), user (comments, global annotations) and complementary graphics.

With Author role: check your algorithm looking for data out of the ordinary by exploring the band of variation of the histogram. You can copy the context components (dictionary, histogram, intervals, pareto / rescaled range) to the internal clipboard so that you can paste it into the Stack or Desktop

With Admin role: check the KPI context the procedure name, formula and algorithm if you need to reuse or create a new KPI

With Developer role: you can use Morph halo to manipulate context components

Fig. 30. Context Inspector

Fig. 31. Data panel context

Library: User defined measures

It is recommended that each organization defines its own plan of measurements, can even begin to work with factory available indicators. The implementation of this plan will be made by any of the following methods available, depending on the nature of the specific indicators required:

Reuse of filters and automatically issue tracker tools (for example Redmine, JIRA) gadgets available as indicators of curved or pie chart (relatively simple indicators derived from data from the incident management tool)

Auto-generation of indicators in the measurement repository from incidence rates and specific fields defined in incident management tool. The application can generate the necesary code for standard indicators of effort, number, and percentage, both curves indicators and group type (pie and bar types)

Coding of new indicators from the measurement data base (or adding the necessary attributes of existing sources or new tools). The user has a high-level language and analytical capabilities created by IMPROVEKIT called IHDSL. This language is similar to SQL, but also to allow use in conjunction with the Smalltalk object-oriented language, contains specific operators for data analysis, for example, search for particular signals. The runtime support and debugging, syntax highlighting, complete fields are a real development environment. The data sources are specified in SQL (if desired, the user or IMPROVEKIT on demand, may also encode indicators in SQL)

Fig. 32. Editor user-defined indicators

Smart categories

They consist of a form of indicator and project filters, defined by the specific IHDSL domain language. In this way, you can define dynamic categorizations according to whether a condition (true, false) is met. The categories can be accessed from the browser or from the menu of scripts (bundles).

For example, you could define a smart category to explore indicators / projects that are similar or different in some respect.

The following smart category (qi) lists the indicators with a certain coefficient of quality:

Kpi qi <0.7

This other smart allows filtering projects of important (fuzzy) performance:

prj performance = 'muy alto'

Inspector

You can find more information on the selected indicator by this panel, so to have in one place the (numerical and graphical) analytical information you need: average values, boundaries, graphics and histogram intervals, pareto (for certain types of indicators)

Detailed View

You can explode each point, bar or portion of any indicator to check the source of its information. To do this, double click on the item to explore and will open in a new window query for the data source involved.

Use this specialized viewer to see more or fewer fields, highlight special information (eg out of range), copying and saving, and if you use an issue tracker, open each issue through a double-clicking on its identifier.

Note: in order for you to have a comprehensive information context, information is displayed (with the aforementioned possibility of filtering) of the data source for the period involved

Fig. 33. Example of an indicator's data source

Working with indicators

View indicators

With the Repository Browser tool you can perform operations with the set of indicators present in the organizational repository of measurements, such as search, filter, query, analyze values, update, etc..

The "View" menu allows you to control how to access the repository. The options let you see the indicators by identifier or by its descriptive name, group them by category (such as Effort, Quality, Productivity, etc..) Or its related process (such as Project Management, Sales, etc..) view externally defined indicators (in Redmine, or JIRA like tools) and filter the list of projects (by category or qualitatively by "performance").

The list of indicators is decorated to give you more information:

Underlined: those indicators whose number of records exceeds the set threshold (see View menu option)
Bold: user-defined indicators (with the application or externally as JIRA filter or Redmine)
Italic: indicators as number of records is not significant (<= 7) for statistical analysis
Strikeout: inactive indicators

Maintain indicators

You can define external indicators (in Redmine for example) to be viewed in conjunction with other indicators in the tool Repository Browser.

Moreover, since this tool you can turn indicators no longer use, or add a specific indicator as given project.

To add new indicators have several options (see "User-defined Measurements"), one of which is self-generate through the menu option

Search Indicators

You can look for indicators in organizational measurement repository in several ways, according to each need:

Navigating with the tool Repository Browser (select a category, a process, or a smart category)

Using the "Go" menu: an indicator directly selecting one bookmark or recently visited. This menu allows the ' relative navigation ' ie, to search according to the current context according to the selected indicator or category other similar (or different)

Searching for text using the search bar (top right). This project can also search by name, within the repository or dashboard

Fig. 34. Search metrics in a tool by text

View Indicators

You can consult an indicator using the data repository tool by selecting the indicator and a project. By default is selected project data, grouped by "phase". You can also consolidate information for the selected project category (for example size). You should always select the project, since it is used for both searches per project to consolidate information by type of project. Depending on the indicator, you can configure the query search more parameters: weekly detail instead of phases of the project, detailed grouping by project category ("baselines") and trends (per phase and project category).

You can view the indicator data in both tabular and graphical format, in order to analyze the variation in the data over time. There are three types of graphs according to the type of indicator:

Curves: show the variation of the data around the average (+-3sigma)

Pie: shows the percentage share of up to five data clustering (available average, maximum, and final series)

Bar: shows the values of a variable amount of data clustering (viewable vertically or horizontally average, maximum and final series)

Fig. 35. Bar chart (vertical)

Fig. 36. Chart circular type (three series)

View same indicator for other project's group

You can quickly see the same indicator at selected but for another project category using the menu option "Go". This will not need to select any project for organizational or baseline clustering corresponding to a category of projects (for example, by size) and is useful for comparing the same indicator in different projects category

Compose a Story

To monitor processes is useful to focus visually information easily accessible. To do this, you use the Dashboard tool, which allows you to perform data analysis of a set of indicators, and other activities as record "baseline" (or comparison), share data with other users and the observations made. Control panels must be constructed selecting indicators of measurement repository. To assist in this task, you have the following options:

Add the selected indicator (listing the name)

Add indicators selected category (eg, all "Sales")

Add indicators that show any sign of variation in the data

An option to add flags with a particular signal variation (5 types are supported standard statistical signal variation)

Fig. 37. Building a Dashboard based on change patterns

Working with charts and numbers

Share information

You can share the graphics and table with numeric values by exporting to image files and external text using the menu option "Share" or send them by email (by default, to the project leader)

Filter indicators data

To analyze subsets or exclude certain values, for example due to particular circumstances, for error in data capture, or peaks that should ignore in the analysis, you can filter the points curve graphs excluding particular points to note the beginning or end of the series via the context menu (right) pointing to a point on the graph.

To perform an analysis with a focus on possible underlying processes, given the behavior of the series, you can filter the type of signal values, particularly those out of range (dominant signals) to exclude or include only those that show a possible process which causes a downward shift, or the rise, the average (weak signals).

This is convenient for signs of change in the indicators of group type as those represented by graphs pie and bar type: for this format can be displayed in a new window curve total values (sum, average or as the case , groups) or a particular group of the series.

From the "View" menu you can filter from an occurrence date of an event corresponding to the "Global Context"

You can keep the filters from-to-exclude from the morphs of the desktop synchronized

Fig. 38. Manipulating the data serie example

Identify values

Relate the points on the graph and number table will help you understand the whole information in both formats. Point to any point on the graph with the mouse to display information related to it, or request identification number table, which is valid for position values, refer to as the minimum, maximum and average. In long series can find projects, periods or particular values in order to direct access to the measurement

Annote a (curve) chart

You can annotate graphics with text to indicate comments, that will be persistent when saved the work environment. You can return to viewing (or hide) through contextual menu (right) graph panel. Other automatic annotations allow you to:

Identify moments or events of the "signs" in the data: periods of the points "dominant"

Analyze data in a qualitative way: legends qualitative type natural language ("very", "little", "pretty", etc..)

Fig. 39. Annotation example

Add series to a chart

To visualize relationships and possible correlations between indicators, you can compose a graph of curves with series of one or more indicators (indicator is usually the same, but you can use any indicators curve). To add an extra series to an indicator, you must first have open the base graphic in a new window (or a page of the scorecard) and by using the chart context menu (right click on it) select "add extra series". You can add as many additional sets as you want (each range is displayed in a different color).

Add a polymetric serie

You can display in graphical form relationships between two indicators, for example, add information from a duration indicator to a efforts scorecard to see the relationship between two variables. This way you compose with a curve type indicator, one polymetric, which takes information from two indicators. You can see the extent of the second variable depending on the size of the related point, according to their value

Fig. 40. Relating two indicators in one chart example

Analyzing data distribution (Pareto)

You can display concentrations in graphic data to determine the most important factors of a, typically large set of factors. To do this, use a chart type "Pareto" graph contains bars and lines where individual values are represented in descending order by bars, and the cumulative total is represented by the line. In quality control, the Pareto chart often represents the most common cause of defects, type of defects with higher occurrence, or the most common cause of customer complaints

Fig. 41. Viewing values grouping and concentration example

Analyzing data buckets

When variables will take a large number of values or the variable is continuous, you can graphically display data clusters in class intervals, allowing you to see the count and repetition frequencies of data easily

Compare using diffuse signals

Diffuse signals are qualitative variations in the data of one indicator relative to another set of data that are calculated using fuzzy logic. Each signal corresponds to a variation pattern of a set of fuzzy values (high, low, medium and intermediate) that can be displayed with the option View Tags | Diffuse Signals.

By default, the set of data to which they are compared are organizational aggregates for the same project type as the selected one. Alternatively, you can select another data panel (with the same indicator as the one selected) from the desktop or the dashboard as comparable, so that your data will be taken as the basis of comparison (for example you could qualitatively compare the values between two projects, set of projects , Or the same project in different periods)

Configure charts colors

For best viewing (and considering special visual capabilities), you can define a range of colors for charts of indicators, from the selection of a basic color range. On the other hand, preferences menu allow you to set default data signals color

Data Panels

Like graphics, data grids can also be displayed as morphs on the desktop. The data can be the values of a data panel that accompanies the graph of an indicator, or query results (written by the user in IHDSL or as a result of exploring a particular point in detail).

The morph is manipulated by using the halo. The menu icon allows access to a set of actions to manipulate the data such as sort, group, and filter. In the latter case you can activate / deactivate the filtering synchronization by date (if applicable) with the rest of the morphs of the desktop.

From the menu of the morph you can "Emphasize" values outside the normal statistics, for example to identify load errors or to detect possible causes of the variation. Values can be plotted in different ways (see next section).

The bottom / center icon of the morph allows you to open an IHDSL script editor so you can write queries with the data to open other grids / graphs or generate new metrics stored in the measurement repository. If the IHDSL code was saved to an external file, the data grid it generates can be kept up to date by selecting the option Refresh.

Fig. 42. A data panel

Charts

From the halo menu of a data grid you can graph the value of a selected numeric field. The generated graphic opens as another morph on the desktop, retaining the context (the data itself).

There are different options available:

Series: x-y line chart
Regression: graph of x-y lines with trend curve trend (best fit). The trend curve can be extrapolated
Limits: x-y chart with limits and statistical signals. You can open a histogram of frequencies (sigmas)
Log: graph x-y log-lin (logarithm-linear)
Re-Scaled Range: v-statistic graphic with hurst coefficient and fractal dimension
Bars: bar graph (sorted)

Fig. 43. Charts of a data panel