The Visualization for this project focused on topics in the financial sector. Finance can be a complicated or overwhelming topic for most people causing them to lose interest or avoid dealing with it in general. It is an industry where the application of visualization goes a long way in helping internal and external stakeholders gain a better understanding of the dealings of different projects. This project focused on investments in finance, specifically hedge fund strategies. The visualization covers the performance of top performing strategies from 2006 – 2013. The goal of the visualization was to reveal an overview of the efficiency of the performances over time and in relation and comparison to other strategies.The Hedge Fund industry publishes the annual income of the funds and fund managers on a yearly basis, those outside of the industry are always curious to these results as they provide insight into sectors of the society doing well financially and possible avenues for personal investments. To understand the data published, the average person might not need to know the nitty gritty of details but a good visualization can go a long way in conveying the right kind of information to inform interested investors. This visualization looks to give a glimpse into different strategies and the amounts invested into them. This information is usually used when Hedge Funds or banks in some cases are looking to review the performances of potential acquisitions. It is traditionally a standard way in the finance industry for data analysts to produce information across time. The desired end product would be charts that give clear information plotted in a financial industry acceptable and professional manner.
The process began with speaking with data and finance experts to gain an understanding of what current practices in the industry are and a better understanding of how these industries react to the emergence and application of visualization. The first interviewee was a data scientist, the intention of the interview was to gain an understanding of what type of data is used for financial visualization in finance. The interview revealed a lot about algorithms used to manipulate the data and the importance of the right data to produce right results. At the end of the interview one of the most important takeaways was that a financial visualization needs to have a purpose to what it is revealing. A visualization created might look beautiful but has to serve a purpose and produce insights no matter how straightforward and boring the information might be, failure to achieve this means the visualization has no meaning or value for an organization.
The second interview was with a financial professional that deals with visualizations for valuations for clients and his company of work. In the interview he revealed the importance of visualizations as they help clients get a quick grasp and overall view of financial information. One of the important take aways mentioned in the meeting was the importance of keeping the flare or designs of the visualization simple and straightforward, as finance from his experience is more of a traditional practice. Leaders in the industry like the information to be kept to the simple and to the point with the need for flashy or cool designs of the visualizations integrated only when needed and not in a distracting way. He said this practice also reads as being professional, which is a major plus if you look to win the bid for an acquisition deal against other competitors.
The third conversation held was with an visual illustrator, explaining the concept of the project and seeking any visualization tips or insights that might come from the creative fields. While most people assume art and creativity can be the farthest thing from finance and numbers, effectively communicating in a visualization is still dependent on following basic color guidelines and principles.
Rationale for design choices.
Based on the information gathered from my interviews, to guide the visualization I took to keeping the design straightforward with the hope of making it a visualization that revealed a clear overview of the information. Also I took into consideration what means would best work in having broken down views of the strategy performances of the amounts across time. The data consisted of 9 different hedge fund strategies namely (Activist, Arbitrage, Macro, Multi-strategy, Event Driven, Fixed income and Directional) the market share amounts and years. To assist in the creation of the visualization, the data was cleaned and transposed in Openrefine and saved as a csv file.
Once the data was ready, it was imported into Tableau public. Once in Tableau the data which comprised of the top 100 hedge funds that were grouped into 9 strategies, was then visualized multiple ways. The data was then visualized first to see which of the charts gave the best reflection of the performances across time. The initial design choice in starting off the visualization was to create a clean line graph, the idea was to create a stripped down version of the information separate and clear and to see if this would reflect the desired information completely. The line graph visualization did reveal the flow of performance over the year but was lacking in a strong reflection depth of the strategies, so the area amounts were filled in to help add a richer depth.
Once area was added a stacked area chart was created to create a better comparison between strategies.
The stacked area chart helps create a mental picture of the overall amount has compared to the neighbouring strategies.
It also gives a clearer single picture of the ebbs and flows across time in one snap shot. Trend lines were added to the line graphs to reveal any trends in the plot. Also a horizontal line graph was plotted to breakup the format seeking other insights that could be revealed in the data.
In looking for a candidate to best reveal certain insights of the strategies, the bubble plot is suggested to be a useful one as it possess capabilities to zoom in and reveal detail on demand of the Information. In this Visualization, the bubble plot was used to show the concentration, distribution and overview of performances of the strategies over the years as well. The bubble plot does a good job at showing a more granular result at each point relating to each year.
In comparing the strategies within the same year some strategies have larger bubbles and others have smaller sizes as small as dots, these chart helps for conversation around what happened within a specific year. This is also the case when comparing the strategies across the years.
Findings, both of the visualization and of the UX research:
In looking to analyze the results of the Visualization a few areas were looked at and discussed below:
Overall the structure of the visualization was intended to be simple focusing on the data in use and amounts. The main focus was to show amounts overtime and secondly performance compared to other external strategies. The top 100 hedge funds across resulted in a list of funds that were grouped into 9 strategies. Then each strategy could be compared internally overtime as well. The display of the data works by selecting a visualized strategy segment (for stacked area or bubble for bubble plot) within the chart the tag pops up to give information like the amount in the strategy and year. This method should one further drill down, can go on to reveal multiple insights from these findings, like who the fund manager was at the time and what were the investments made, these answers go on to give valuable historical investment information.
How did scale work:
The scale of the charts were able to convey the overall amounts, as it wasn’t dependant on size thanks to the label tags included in the charts. In Dashboard 2, Bubble Plot with the highest performing Hedge Fund Strategy, Activist excluded, the Activist which had a huge outlier bubble of over 54 million was excluded from the layout to help give a closer look at the other strategies.
How did the layouts orientation work:
Different layouts helped reveal different areas of focus. As mentioned earlier the area stacked chart layout helped in giving a general overview of the entire chart. The horizontal bar chart gave an overview in regards to measurement comparison and the bubble plot helped give an overview of the charts performances internally and externally across the years.
Though there wasn’t an extensive drill down via the bubble plot, it enabled us to zoom in and have a clearer view of the performances in each year.
How did order alignment work:
The alignment of the strategies against the years in the rows and columns helped in the display of performances overtime. Also, the alignment of the strategies against each other helped in giving a snapshot of comparison of the information in a straightforward and non-confusing way that allows a user figure out the information without needing an instructor. This approach which is common in the finance industry helps make for an easy comparison and overall read of the charts. Then when presenting to clients the data analyst when requested zooms into a specific strategy for drill down or more granular insight. However for the purpose of this visualization, only a top level view of segments was created.
UX research Findings:
The user experience played a major role in guiding the final direction and nature of the product deliverable. Based on the feedback from interviews, understanding the nature of the finance industry and how hedge fund strategy visualizations are used was insightful and important. Creating the visualization had more to do with the culture and type of user than ability and flashy designs. While a cool and nice design is good, in finance it is desired only as relevant as the features would help efficiently inform the users of this traditionally humdrum numbers focused industry. Since most visualization of this nature are presented for the purpose of comparison of segments, a strong focus was placed on visualizations that can depict clear representations in an intuitive way to a non specialist reviewer.
Since the visualization was straight forward just reflecting segments for strategy returns, the color choices selected were intended to add a little life to the layout to avoid a dull portrayal of the information. Bright colors also aided in providing contrast to the categories with the intention of helping the visualisation be easier to read.
Which strategy had strong returns?
The Activist, Macro and Arbitrage had strong performances with Multi-Strategy and tracking following behind. In 2012 the Activist had the strongest returns of over 54 million making it an outlier in the bubble plot.
Which strategy did strong across the years?
Based on the visualization while the Activist strategy acquired the most amount of the market share of investments. Steady performances appeared to be happening in the Macro and regional, despite the fact these had lower shares of the market. Again while the income wasn’t that hight different investors seek different criterias for their investments and this insight proves helpful. This is impressively revealed via the visualisation as opposed to looking at this on an excel sheet.
The final result of the visualization helped showcase trends and reveal hidden behaviors that could easily have gone un-noticed just looking at just a regular excel sheet. While the final visualization gave clear indications to how much each strategy amassed during the 7 year investment period, there are always other things to consider that could have been external factors to the results. This visualization gives an overview of the kind of information that is presented to the investor but is in no way an exhausted measure of the total ways this is done in the financial industry. Most standard ways of visualizing data require an overview, an ability to filter, zoom in on key segments and present more granular details by on demand. Future visualizations would do well in showing other kinds of industry standard analysis which would require more breakdowns or drill downs of the charts. Other recommendations could be to use R to better instruct the data structure via algorithms for higher optimization of the data set.