Examples of people wanting to compare variables - salspaugh/DataFramer GitHub Wiki

memo 28, dataframer:

"It's hard to say whether the effects I see in the data (only 8 diverted flights) are representative of that time of year or some airport -- for example Atlanta seemed over-represented compared to O'Hare."

memo 41, tableau:

"My next step for this examination would have been to compare those delays relative to the total number of flights coming in and out of the airport...Overall there seemed to be a high correlation between airports experience both arrival and departure delays. But again, my next step would be to compare that as a ratio to the total number of flights coming in and out of the cities."

memo 42, dataframer:

"I also find it challenging to really come to any conclusions on my questions (or therefore to evaluate if I should keep or reject them) since I can't really correlate the columns to one another. I know there are a high number of single birds struck, for example, and that there was a high number of no damage of military planes. Are these correlated? For example, maybe all the no damage military planes were struck by 2-10 birds. The result could be interesting if I could compare."

memo 42, dataframer:

"For my final question I wanted to see if precipitation correlated with accidents at certain times in the flight path. I was surprised just looking at the precipitation data alone that most accidents happened in rain or fog. I would expect snow would have a high correlation of accidents too. But I guess rain and fog is generally just more common than snow. Or it's too cold for birds to be out."

memo 52, spreadsheet:

"We could compare how well various types of airplanes resist hits with wildlife and how others potentially have much larger downtime and damage."

memo 59, dataframer:

"My questions are hypothetical correlations between delays and flight properties. Are flights more likely to be delayed when arriving to airport X? Do delays better correlate with start or end airport or start or end city market? Are delays uniformly distributed across the days of the week? How do busy periods (Christmas, Summer) affect delays? Are flights more likely to be delayed when departing from airport X? Are delayed most likely incurred before take off, in flight, or after take off while taxiing? Are some airlines more frequently delayed than others? By finding what correlates with delays, we can hope to predict and perhaps even fix the underlying causes of delays."

memo 64, dataframer:

"Is a particular aircraft type more prone to strikes? Is a particular aircraft type more likely to be damaged by strikes? Very little effect data. Is a particular area more prone to strikes? Why are so many more of the strikes in approach than in any other stage? I assume a null in precip means the sky was clear? Why are particular runways more prone to strikes? I wonder whether there's a relationship between the species hurt and the type of aircraft, or the species hurt and the position in the flight at which the strike occurs. There appears to be a yearly cycle in the frequency of incidents, with peaks in the late summer. I wonder whether other features correlate with these peaks. Why does American Airlines have so many incidents? It seems this dataset will not give us insight into the ratio of flights an airline runs to strikes it experiences. I'm also curious about business and military being so high. Similarly, I suppose this dataset cannot reveal whether a helicopter is more likely than an airplane to experience a strike on any given flight, since we do not have information about the number of flights by the various crafts that do not experience strikes. Why does Florida have such a high incidence of strikes? Its size does not seem to account for it, since Texas and California have lower rates of enroute strikes. Of course, there is actually very little data on this, since there are so few entries for enroute. I wonder what features of strikes tend to vary from helicopters to airplanes. What is Lihue Airport, and why does it have such a high number of strikes? I wonder if there is a relationship between number of birds struck and amount of damage, or number of birds seen and amount of damage. I wonder whether there are few reports of strikes in snow because there's little snow, or because strikes actually become less likely in snow."