> For the complete documentation index, see [llms.txt](https://undp.gitbook.io/performance/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://undp.gitbook.io/performance/methodology/values/transparency-index.md). # Transparency Index ## Transparency Index This is the global indicator. ### Introduction to Indicator {% embed url="" %} ### Calculation of Scoring. Publish What You Fund Score = Indicator Score The traffic light indicator methodology for this indicator is: * Green = 80+ * Yellow = 60+ * Red = <60 {% hint style="info" %} This is aligned with the Very Good / Good / Fair scoring at Publish What You Fund {% endhint %} ## Transparency Dashboard Performance. {% hint style="info" %} This is a draft and under testing {% endhint %} This is the indicator for Country & Bureau levels. ### Available Data Downloading the dataset from the UNDP Data Warehouse: * UNDP\_PROJECTS data - List of all unique projects in UNDP Columns in UNDP\_PROJECTS data * hq\_co: Headquarters (HQ), Country Office (CO), or RC. * bureau: The name of the specialized unit or division responsible. * rollup\_ou: Organizational unit's code. * rollup\_ou\_description: Organizational unit's name. * PROJECT\_ID: Project identifier. * PROJECT\_NUMBER: Project number. * PROJECT\_NAME: Name of the project. * PROJECT\_DESCRIPTION: Description of the project. * ATLAS\_AWARD\_NUMBER: Atlas award number. * ATLAS\_AWARD\_DESCIPTION: Description of the Atlas award. * BUSINESS\_UNIT: Business unit identifier. * ORGANIZATION: Organization involved in the project. * DEPARTMENT: Department associated with the project. * START\_DATE: Start date of the project. * COMPLETION\_DATE: Completion date of the project. * CLOSED\_DATE: Date when the project was closed. * PROJECT\_TYPE: Type of project. * PROJECT\_TYPE\_DESCRIPTION: Description of the project type. * PROJECT\_STATUS: Status of the project. * PROJECT\_MANAGER: Project manager's name. * PROJECT\_MANAGER\_EMAIL: Email address of the project manager. * IMPLEMENTING\_PARTNER: Implementing partner code. * IMPLEMENTING\_PARTNER\_DESCRIPTION: Description of the implementing partner. * IMPLEMENTATION\_MODALITY: Implementation modality code. * IMPLEMENTATION\_MODALITY\_DESCRIPTION: Description of the implementation modality. * PROJECT\_ORIG\_TEMPLATE: Original template of the project. * PROGRAMME\_FUNDING\_FLAG: Flag indicating program funding. * GEF\_GCF\_PROJECT\_FLAG: Flag indicating GEF/GCF project. * Other\_References\_Value: Other references value. ### Calculation: 1\. Define the rollup\_ou\_description to filter: Filters the DataFrame df\_UNDP\_PROJECTS by selecting only the rows where the value in the 'rollup\_ou' column matches the value stored in the variable target\_rollup\_ou (note: which is 'ALB' in this case as an example). 2\. Calculates the count of projects for each unique project status in the dataframe if count>0. The statuses are: * Financially Closed * On Going * Operationally Closed * Submit for Operational Close * Submitted for Financial close 3\. Visualization - Grouping projects by project status and counting the number of projects in each status. 4\. Filter data by specified project statuses: Select rows from the dataframe where the 'PROJECT\_STATUS' column matches the specified status values: 'On Going', 'Operationally Closed', and 'Submit for Operational Close'. Then remaining visualizations are done using this filtered dataframe. 5\. Visualization: creates a horizontal bar chart showing the number of projects by project type, with the bars colored in sky blue except for the bars representing projects with null Project\_Type, which are colored in red. Grouping projects by project type and counting the number of projects in each type - First, count the occurrences of each unique project type in the 'PROJECT\_TYPE' column. Then resets the index and renames the columns to 'Project\_Type' and 'Count' respectively to prepare the data for visualization. Then, filters the dataframe to select rows where the 'PROJECT\_TYPE' column is null and counting projects with null Project\_Type. After that, adding the count of projects with null Project\_Type to the dataframe. Then created a bar chart including the rest of the project types. 6\. Visualization: Counting the number of projects with and without descriptions - counts the number of non-null values in the 'PROJECT\_DESCRIPTION' column and then calculates the number of projects without descriptions by subtracting the count of projects with descriptions from the total number of projects. Then created a bar chart and annotated each bar with its respective count. 7\. Visualization: Counting the number of projects with and without PROJECT\_TYPE - calculates the number of non-null values in the 'PROJECT\_TYPE' column and then calculates the number of projects without a 'PROJECT\_TYPE' by subtracting the count of projects with 'PROJECT\_TYPE' from the total number of projects. Then created a bar chart and annotated each bar with its respective count. 8\. Visualization: creates a horizontal bar chart, annotates each bar with its respective count - Computes various statistics about projects on such as: * Total number of projects * Projects with less than 10 characters in project name * Projects without description * Projects without implementing partner * Projects with description less than 80 characters Transparency Scoring for CO - Methodology: Step 1: Make sure having a filtered dataframe (rollup\_ou = \[mention CO], also, PROJECT\_STATUS'].isin(\['On Going', 'Operationally Closed', 'Submit for Operational Close'). Step 2: Penalty Factors for More Severe Penalization (1 to 5): Define penalty factors for each indicator to adjust the severity of penalization - * Penalty factor for projects with names shorter than 10 characters = 1 * Penalty factor for projects without a description = 2 * Penalty factor for projects without an implementing partner = 4 * Penalty factor for projects with descriptions shorter than 80 characters = 3 * Penalty factor for projects without a project type = 5 Step 3: Count of All Projects : Calculate the total number of projects for normalization purposes. Step 4: count the occurrences of projects in each noncompliance case: * projects with names shorter than 10 characters. * projects without a description. * projects without an implementing partner. * projects with descriptions shorter than 80 characters. * projects without a project type Step 5: Determine total number of projects for each indicator. Step 6: Scores Calculation: Calculation of Percentages: Calculate the percentage of non-compliant projects relative to the total number of projects for each indicator. Penalty Factor Application: Apply the penalty factors to the percentages of non-compliant projects instead of raw counts. This adjustment ensures that the severity of penalization is proportionate to the extent of non-compliance relative to the total project count. Scoring Calculation: Calculate the scores based on the adjusted percentages of non-compliant projects multiplied by 20. This scaling operation ensures that the scores fall within a range of 0 to 20. Finally, individual scores for each indicator, along with the total score out of 100 for the CO, are displayed. Note: With each indicator having a maximum of 20 points and 5 indicators, the maximum possible total score would be 100 points.