Connecting the Dots: The Power of Network Graphs in Identifying Business Experts
In the fast-paced world of business, it is critical to attract and retain top talent. Every organization aims to have a team of experts who are knowledgeable, dedicated, and hardworking. These individuals help companies achieve their ambitious business goals and drive growth.
To identify and recognize such experts, Exness recently conducted a non-anonymous survey of its employees. The survey consisted of seven dropdown lists of employees, and the results were used to identify experts in various areas. The goal was to involve these experts in decision-making processes, empower them to contribute, and reward them accordingly.
Recognizing and rewarding employees who contribute to achieving business goals is essential to the success of any organization. Employees who feel valued and appreciated are more engaged, motivated, and committed to their work. This, in turn, can lead to increased productivity, better job performance, and ultimately, improved business outcomes.
For a more detailed exploration of this topic, I encourage readers to delve into my colleague’s Alexander Nezhelsky insightful article From Employee Voices to Expert Stars: Uncovering the Most Valuable Assets in Organizations. He provides a comprehensive account of the business need, the survey process, and the algorithm used to determine the weight of each employee’s contribution. It is an excellent resource for gaining a deeper understanding of the subject matter at hand, and I highly recommend giving it a read.
In this article, we will explore how network graph technology can help organizations identify and recognize their top experts. We will also delve into the technical details of using Gephi and Tableau to create network graphs using data obtained from the survey. This type of visualization shows how employees are interconnected through the use of nodes and link lines and helps identify key experts across various areas of the organization, thereby gaining valuable insights into their strengths and areas of expertise.
Using Gephi to Get X, Y Coordinates
To create a network graph in Tableau, we first prepare the data and assign X and Y coordinates to each employee. This can be done using Gephi, an open-source software for network analysis and visualization.
Assigning X and Y coordinates using Gephi involves two main data files: a Nodes file and an Edges file. The Nodes file contains information about each employee, which is populated with two columns Id and Label (Name). Be sure there are no duplicated nodes!
The Edges file contains information about the connections between employees, such as who they nominated as business experts. We then need to create two columns called ‘Source’ and ‘Target’ and populate them with the appropriate employee IDs. This will create the necessary links between employees in the network graph. Also, a division field could be added to make partitions for employees.
Once we have the Nodes and Edges files, we can import them into Gephi and use various algorithms to assign X and Y coordinates to each employee.
Import Nodes.csv:
Import Edges.csv:
Next, we need to calculate a PageRank using an in-built algorithm. After assigning X and Y coordinates, we can use PageRank as size and color for the nodes to show the influence of each employee. We can also use the division as a colored partition for the edges to show the connections between employees in different divisions.
One popular algorithm for graph visualization is the ForceAtlas 2 layout, which uses a physics simulation to place nodes in a way that minimizes edge crossings and maximizes node-edge distances.
Once we have the layout we want, we can export the coordinates as a JSON file (File => Export => Graph file => *JSON) and import it into Tableau. This JSON file will contain the X and Y coordinates for each employee, which we can use to create the final network graph and dashboard in Tableau.
Data Preparations
Data preparation is essential in creating a network graph in Tableau. In this section, we will discuss the process of converting the JSON Coordinates file into csv, duplicating data from the Edges file, adding ‘Direction’ and ‘Location’ columns to the data, and adding additional rows as self-references for employees who did not participate in the survey.
First, we need to convert the JSON Coordinates file into CSV that could be easily imported into Tableau. This was done using a Python script. Once we have the CSV file, we can use it to assign X and Y coordinates to each employee in Tableau.
Next, we need to duplicate the data from the Edges file to create the lines in Tableau. To do this, we need to copy and paste the data from the Edges file into a new Google/Excel sheet and UNION it, i.e. duplicate it down to the original data. We also need to add the ‘Direction’ and ‘Location’ columns to the data. The ‘Direction’ column is added to indicate the particular part of the data — Source (first) or Target (second unioned part). The ‘Location’ column is added to indicate each employee’s (node) location on the graph. The Location column joins the Coordinates file and Survey file in Tableau. So, for the initial part of data Direction=’Source’, Location = Source Name. For the second copied part of data Direction=’Target’, Location = Target Name.
Finally, we added additional rows to the data to account for employees who did not participate in the survey. These rows acted as self-references for employees who did not have a Source row in the Edges file. This allowed us to include all employees in the network graph and see their relationships with other employees.
Building a Network Graph in Tableau
Once the data is ready, the next step is to build the network graph in Tableau. First, we connect data and create a Tableau data source. So, we need to connect to the Survey CSV text file and join it with the Coordinates CSV file (be sure these files are in the same folder) using the Location-related field.
Next, we should assign the geographic role to the X and Y fields. This will allow Tableau to recognize the coordinates as geographical points and enable the creation of a map.
Then, a calculation called the ‘Path’ is created. The Path calculation will connect the nodes based on their X and Y coordinates, creating lines representing employee relationships. The Path calculation is used to display the lines on top of the map.
To create the circles representing the employees, the data is duplicated and added as a Dual Synchronized Axis. Rows and columns are then swapped to move the circles to the correct position.
To size the circles based on the weight of the connection, the weight field is added to the Size shelf. The division is added to the Color shelf for the lines, allowing them to be color-coded by division.
Overall, building the network graph in Tableau allows for a more interactive and visually appealing way to analyze the data. The use of color and size helps to identify patterns and trends in the data, making it easier to identify influential employees and their relationships.
Dashboard Insights
A network graph is a powerful tool for visualizing complex data and identifying patterns and relationships that may not be immediately apparent in traditional forms of data presentation.
The dashboard created using the network graph provides a wealth of insights into the expertise and connections of employees in different divisions and locations. By selecting a particular division, it is possible to see who the selected experts are and how they are connected to others in the organization. This information can be particularly valuable in identifying key experts who are deemed crucial to the success of the business according to the opinions of the employees.
Additionally, by selecting non-managers, it is possible to identify employees who are influential in specific areas or who have developed specialized expertise. This information can be used to guide decisions around promotions, career development, and training opportunities.
By selecting a particular location, we can see which employees are most connected within that region and how their expertise is distributed. This information can be particularly useful in identifying opportunities for cross-functional collaboration and knowledge sharing.
Furthermore, the dashboard allows to see who selected a particular expert and how that expert ranks within their division or location. This information can be used to recognize and reward employees who contribute to achieving business goals, as well as to develop strategies for knowledge transfer and succession planning.
Finally, the dashboard allows users to compare the ranks of chosen experts between employees, providing insight into the expertise and influence of these employees across different divisions and locations. Overall, the insights provided by the dashboard can be invaluable in guiding decisions around talent management and business strategy. By visualizing the data in this way, we can identify new opportunities for collaboration and innovation within the company.
Conclusion
The use of technologies like Gephi and Tableau can help businesses identify and recognize employees who contribute to achieving business goals. The network graph created in Tableau provides a visual representation of the relationships and connections between employees, making it easy to identify those who are influential in certain areas and visualize the distribution of expertise across divisions and locations.
Facilitating the identification of esteemed experts according to their peers’ opinions through such tools can help bridge the gap between top management and on-the-ground experts, enabling businesses to effectively recognize and reward employees who surpass their job responsibilities in driving the company towards its goals. Doing so not only boosts employee morale but also motivates employees to continue to work hard and achieve ambitious business goals. By using a network graph, businesses can better identify and celebrate the experts in their organization, creating a more collaborative and inclusive environment. Therefore, leveraging the power of data visualization tools can help companies better identify and reward these valuable contributors, ultimately leading to greater business success.
P.S. I would like to extend my sincere gratitude to my colleague Alexander Nezhelsky for his collaboration on this project, as well as his invaluable assistance in data preparation. His support in handling the data was instrumental in achieving the project’s success. Special thanks to Polina Lukina, and Ilya Lokotilov for their incredible passion and valuable impact.
