Skip to content

Understanding Undertaking Analysis and Similarity Evaluation | by Tushar Babbar | AlliedOffsets | Oct, 2023

Within the realm of undertaking analysis and decision-making, it’s essential to comprehensively analyze undertaking particulars and discover potential alternate options. This course of entails not solely summarizing project-specific information but additionally assessing the similarity between initiatives to uncover priceless insights.

This text delves into the intricacies of undertaking analysis and similarity evaluation, shedding mild on how numerous attributes play an important function in figuring out undertaking similarity. We’ll discover the attributes thought of, their assigned weights, and the step-by-step course of for calculating similarities.

Let’s dive into an in depth rationalization of the built-in course of that features each acquiring undertaking particulars and calculating similarities between initiatives. We’ll break it down step-by-step:

We start by creating an inventory of distinctive IDs (UIDs), that are like undertaking identifiers. Every UID represents a selected undertaking in our dataset. These UIDs are essential for referencing and retrieving detailed details about every undertaking.

Assigning Weights to Attributes:
Within the means of discovering comparable initiatives, we assign weights to totally different attributes to find out their relative significance in calculating the similarity between initiatives. These weights information the calculation of a similarity rating, which quantifies how intently two initiatives align by way of these attributes.

Attributes and Weights Used for Similarity Calculation:

  • Continent(0.8) and Nation(Weight: 0.7): Tasks in comparable geographic areas could have similarities as a result of native elements.
  • Registry (Weight: 1.5): The registry the place a undertaking is listed can point out similarities by way of rules and trade.
  • Sectors(Weight: 1.0) and Subsectors (Weight: 1.5): Tasks categorized in comparable sectors and subsectors would possibly share comparable aims or traits with heavier weightage to the undertaking sub sector.
  • Methodologies (Weight: 1.0): Comparable methodologies utilized in initiatives could recommend frequent practices or targets.
  • Area (Weight: 0.5): Throughout the similar nation, the geographic area of a undertaking can affect its attributes and efficiency.
  • Undertaking Acreage (Weight: 0.5): The scale of the undertaking by way of acreage could be a think about similarity.
  • Measurement (Weight: 1.0): The general measurement or scale (Micro, Small, Giant) of a undertaking is taken into account for similarity.
  • Undertaking Exercise Degree (Weight: 1.5): This attribute displays how lively or engaged a undertaking is. Additional particulars on how this attribute is derived could be discovered right here: VCM Liquidity Index.

Sure refinements have been launched to reinforce the robustness of the Undertaking Exercise Degree. When assessing the exercise ranges of two initiatives, a particular method is employed. If the distinction between their exercise ranges is exactly +1 or -1, a weighted aggregation mechanism comes into play. Within the case of a +1 distinction, the load is elevated by 0.2, elevating it to 1.7 from its authentic 1.5. Conversely, when the distinction is -1, the load undergoes a discount of 0.2, leading to a weightage of 1.3 as a substitute of the earlier 1.5. This adjustment is made to make sure that initiatives with exercise ranges intently resembling the in contrast undertaking don’t lose significance, acknowledging their similarity in nature..

With these weights in place, the code then calculates a similarity rating for every pair of initiatives. The similarity rating is derived by evaluating the attributes of the present undertaking in our UID record with the attributes of every undertaking within the dataset. The rating is calculated because the sum of the product of attribute values and their corresponding weights.

Filtering the High Comparable Tasks:

After inputting a singular identifier (UID), it undergoes a complete scan throughout all initiatives inside the database. Every time it identifies a match by way of attributes, a corresponding weight is assigned. These weights are subsequently aggregated, successfully producing a similarity rating. Following this computation, there’s a validation step in place: if the similarity rating surpasses or equals 6, we deem the undertaking as comparable. At this level, we current solely the highest 5 initiatives, ordered in descending order of their similarity scores.


The system selects a uid on this case “Rimba Raya” with uid “VCS674”, and a undertaking for comparability, for instance, “Katingan Peatland Restoration and Conservation Undertaking” with the identifier “VCS1477,” and proceeds to judge the extent of attribute similarity.

On this explicit occasion, the attributes recognized as comparable, together with their respective weights, are as follows:

Continent: 0.8
Nation: 0.7
Sector: 1.5
Registry: 1
Measurement: 1
Exercise: 1.5
Area: 0.5

The cumulative similarity rating, derived from these weighted attributes, yields a complete rating of 7.0.

Subsequently, a situation is utilized to evaluate if the similarity rating meets or exceeds the edge of 6.0. On this case, the situation is glad, resulting in the conclusion that “VCS1477” is certainly just like “VCS674.”

For every UID in our record, we offer a complete report encompassing:

  • Undertaking particulars, together with: Whole Issued Credit, Whole Retired Credit, Whole Retired Credit (Final 12 Months), New Retirees (Final 12 Months), Whole Distinct Retirees, High 3 consumers, and extra.
  • A listing of comparable initiatives, full with their names, similarity scores, and the weights assigned to every contributing attribute.

This detailed report empowers brokers with a holistic view of initiatives, facilitating knowledgeable decision-making and the exploration of potential alternatives inside the dataset.

The ultimate output of this built-in course of is an in depth report for every UID in our record. This report contains each project-specific particulars and details about comparable initiatives, full with attribute weights. Brokers can leverage this complete report back to make knowledgeable selections and take actions associated to those initiatives, harnessing the weighted similarities to evaluate potential connections and alternatives inside the dataset.

In essence, this structured method to undertaking analysis and similarity evaluation empowers decision-makers to navigate undertaking landscapes with higher readability and perception.

For extra data, please attain out to good

continue reading