Deeper analysis of the unusual conversations – part 2

Let’s next take a closer look at the number of films scored by Raja over the past 37 years. Instead of looking at the movies by years, let’s summarize it by years. The number of movies scored by Raja by year is very debatable as there is no accurate record available that is authoritative. I have used databases information available on the internet and I have filled the gaps where I am aware of. The database covers all language films set to music by Raja. While this database may not be very accurate, I find that the best way to use it is to summarize the number of films set to music by Raja over a decade. This helps in:

1. Reducing compensating film count errors over the years
2. Takes out the language aspect of films he has worked on
3. For trend analysis, it is best done on a summary data set instead of the detailed one, as the atomic level data is not deemed very accurate

Decade	#Films composed	Percent
1970	72	8%
1980	421	46%
1990	314	35%
2000	102	11%
Grand Total	909

 

Something interesting happens when we now try to correlate both these data sets. In other words, we do this:

Decade	Unusual conversation occurrences	#Films composed
1970	6	72
1980	151	421
1990	47	314
2000	6	102
Grand Total	210	909

 

When we use a statistical correlation function between Column 2 and Column 3, the correlation coefficient works out to be .91642341.  Any correlation coefficient greater than .7 is considered strong positive correlation between the data sets being compared. Data does not lie!

Let’s try and interpret this at a more granular level – at the year level.

Year	Unusual conversation occurrences	#Films composed
1978	1	25
1979	5	30
1980	7	34
1981	15	33
1982	12	33
1983	20	47
1984	17	54
1985	18	51
1986	21	50
1987	13	32
1988	15	44
1989	13	43
1990	9	46
1991	9	41
1992	7	55
1993	3	42
1994	4	27
1995	7	25
1996	2	12
1997	4	20
1999	2	24
2003	1	11
2007	2	4
2008	1	5
2009	2	18

 

In the data table above, you will find a few years missing. Those are the years, where we did not have any unusual C&R observations. The total of column 2 in this table is the same as before. However, the total of the number of films composed will be lesser as we have excluded a few years where there were no unusual C&R observations. When we use a statistical correlation function between Column 2 and Column 3, the correlation coefficient works out to be .7576654.  Any correlation coefficient greater than .7 is considered strong positive correlation between the data sets being compared.

How do we interpret this?

1. The C&R technique has a very strong correlation to the number of films composed by Raja in a decade. The unusual conversations is used as a proxy for the whole C&R technique. The correlation stays strongly positive, even if the comparison is done at the year level
2. Raja’s highest number of films and the most use of this technique is in the 1980s
3. His number of films composed have fallen significantly in the 2000s and his use of this technique has also fallen, compared the other decades
4. As most average Indian listeners can relate to a C&R technique, especially when it is set to the Carnatic format, they enjoyed it the most during Raja’s 1980s
5. Raja’s increased use of synthesized instruments have either resulted in less use of this technique, or his use has not surfaced as distinctly as it did in the 1980s and 1990s

Obviously, there are many caveats to any analysis:

1. No one technique ever contributes to a composer’s success in Indian film music
2. As most hit songs in the Indian film context have little to do with orchestration (they mainly depend on the melody of the main tune), these results have to be viewed in that broader context

However, Raja is the best orchestrator ever from India. It is worthwhile analyzing his orchestration techniques and their impact. Though the data sets are summarized ones, they show strong correlation to the number of films he has scored in a decade/year. In other words, one can take an extreme view, that ,this technique has a directly correlation to his popularity.  It may be worthwhile redeploying this technique in his future work as it has appeared to have served him well in the past.

 

One can even argue that when Raja introduced harmony and counterpoints to the Indian audience in the 80s, he always provided it along with something known – C & R. It was easy for the listener to enjoy something unknown, when it was served along with something known. It may be perhaps worthwhile for Raja to consider reintroducing C & R as he continues to experiment with newer techniques in the 21st century. If past is any indicator, of future behavior, why not use something that has worked very well in the past?

While this analysis work was the hardest of all the areas I have explored of Raja’s orchestration, it was definitely enjoyable. The data based interpretation of Raja’s work is worthwhile in my view. Hope the readers enjoyed it as much.

Correlation and causation are different. However, correlation in this case, does have a message - highly popular Raja years have a strong positive relationship with the use of this technique. This analysis needs to be done for other techniques over the years to get to a full understanding of the other factors influencing popularity. However, in isolation, with the data gathered in this project, it appears that there is indeed a strong positive relationship between his popularity and the use of this technique.