In [1]:
require(tidyverse)

Loading required package: tidyverse
── Attaching packages ─────────────────────────────────────── tidyverse 1.2.1 ──
✔ ggplot2 3.1.0       ✔ purrr   0.3.2  
✔ tibble  2.1.1       ✔ dplyr   0.8.0.1
✔ tidyr   0.8.3       ✔ stringr 1.4.0  
✔ readr   1.3.1       ✔ forcats 0.4.0  
── Conflicts ────────────────────────────────────────── tidyverse_conflicts() ──
✖ dplyr::filter() masks stats::filter()
✖ dplyr::lag()    masks stats::lag()

Reading the geopanda implementation execution times...

In [2]:
geopandas = read.table("~/RIDIR/Code/Validation/GeoPandasBenchmark.tsv", header = F, sep = "|")
names(geopandas) = c("Timestamp", "Method", "Size", "TotalTime", "AreaTablesTime")
head(geopandas)
TimestampMethodSizeTotalTimeAreaTablesTime
2019-04-23 21:42:41,899GeoPandas 10% 1.97 1.81
2019-04-23 21:42:48,264GeoPandas 20% 5.58 5.30
2019-04-23 21:43:00,281GeoPandas 30% 11.23 10.83
2019-04-23 21:43:22,517GeoPandas 40% 21.44 20.89
2019-04-23 21:43:58,617GeoPandas 50% 35.31 34.63
2019-04-23 21:44:52,760GeoPandas 60% 53.35 52.53

Reading the geospark implementation execution times...

In [3]:
geospark = read.table("~/RIDIR/Code/Validation/GeoSparkBenchmark.tsv", header = F, sep = "|")
names(geospark) = c("Timestamp", "Method", "Size", "TotalTime", "AreaTablesTime")
head(geospark)
TimestampMethodSizeTotalTimeAreaTablesTime
2019-04-23 23:40:47,266GeoSpark 10% 23.66 8.35
2019-04-23 23:41:17,700GeoSpark 20% 27.30 12.84
2019-04-23 23:41:49,609GeoSpark 30% 28.76 15.10
2019-04-23 23:42:17,571GeoSpark 40% 24.77 9.17
2019-04-23 23:42:44,353GeoSpark 50% 23.65 10.15
2019-04-23 23:43:12,160GeoSpark 60% 24.68 10.91

Merging the datasets and preparing the data...

In [15]:
benchmark = as_tibble(rbind(geopandas, geospark)) %>%
                select("Method", "Size", "TotalTime", "AreaTablesTime") %>%
                mutate(TotalTime = as.numeric(TotalTime), AreaTablesTime = as.numeric(AreaTablesTime)) %>%
                group_by(Method, Size) %>% summarise(TotalTime = mean(TotalTime), AreaTablesTime = mean(AreaTablesTime)) 
benchmark$Size <- factor(benchmark$Size, levels = c("10%", "20%", "30%", "40%", "50%", "60%", "70%", "80%", "90%", "100%"))
head(benchmark)
MethodSizeTotalTimeAreaTablesTime
GeoPandas10% 1.972 1.812
GeoPandas100% 147.968 146.644
GeoPandas20% 5.566 5.286
GeoPandas30% 11.298 10.892
GeoPandas40% 21.406 20.854
GeoPandas50% 35.368 34.696

Ploting the execution times for both implementation...

The CA (California) dataset has 5846 polygons in the source and 7953 polygons in the target. Each 10% increase in the size of the dataset is about 584 polygons in the source and 787 polygons in the target.

In [20]:
g = ggplot(data=benchmark, aes(x=Size, y=AreaTablesTime, fill=Method)) +
    geom_bar(stat="identity", position=position_dodge(width = 0.75), width = 0.7) +
    labs(title="Benchmark area_tables using CA dataset", y="Time(s)", x="Size of CA dataset")
plot(g)

Reading execution times for geospark implementation in the full dataset...

In [36]:
data = read_delim("/home/and/RIDIR/Code/Validation/GeoSparkFull.tsv", delim = "|", col_names = F) %>%
        rename(Timestamp=X1, Method=X2, Cores=X3, Nodes=X4, Partitions=X5, Time=X6, Results=X7, Size=X8, AppID=X9) %>%
        select(Size, Time) %>%
        mutate(Size = paste0(Size * 100, "%"), Time = as.numeric(Time)) %>%
        group_by(Size) %>% summarise(Time = mean(Time))
data$Size <- factor(data$Size, levels = c("20%", "40%", "60%", "80%", "100%"))
head(data)

Parsed with column specification:
cols(
  X1 = col_character(),
  X2 = col_character(),
  X3 = col_double(),
  X4 = col_double(),
  X5 = col_double(),
  X6 = col_double(),
  X7 = col_double(),
  X8 = col_double(),
  X9 = col_character()
)
SizeTime
100% 93.216
20% 42.360
40% 47.984
60% 59.570
80% 74.338

Plotting execution times...

The Full dataset has 72693 polygons in the source and 61332 polygons in the target. As reference, the geopandas implementation in the full dataset throws a memory error after ~45min of execution...

In [39]:
f = ggplot(data=data, aes(x=Size, y=Time)) +
    geom_bar(stat="identity", position=position_dodge(width = 0.75 ), width = 0.7) + ylim(0, 100) +
    labs(title="Benchmark GeoSpark area_tables using the Full dataset", y="Time(s)", x="Size of Full dataset")
plot(f)
In [ ]: