In kamilien1/patentr: A Toolbox for Analyzing Patent Data

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.path = "README-"
)

Introducing patentr, the toolkit for patent data analysis in R. The summary md file currently holds all documentation.

The package is aimed at patent agents, lawyers, managers, analysts, and academics who are working on patents. This may be used in a patent landscape analysis, company IP portfolio analysis, or a freedom to operate search.

This is a data processing and reporting tool on patent data sets for patent analysts. The motivation comes from a lack of useful, exportable patent data. patentr builds upon the free data available from Sumobrain.com, Lens.org, and Google Patents, leveraging their data to summarize and analyze patents.

patentr performs four key functions:

1. Data input: Easily import patent excel and csv files from the top patent websites
2. CSV from Google Patents and lens.org
3. xlsx from sumobrain.com
4. Data cleaning: Sanitize patent data and extract useful metadata for custom analyses
5. Clean up important fields such as names, dates, country codes, and kind codes
6. Infer the document type so that you don't analyze the same patent twice
7. Deduplicate data sets and prioritize grants over applications
8. Use the generated Google URL to jump to the patent document or to download claim data using the included httr and XML functions
9. Exploratory data analysis: Explore patent data and quickly visualize important attributes
10. Quickly summarize patent data by relevant columns to get document count
11. View standard histogram, tile, and facet plots of important information
12. Extract claim information for wordcloud analysis
13. Interact with your data on the Shiny user interface
14. Reporting: Export your data as powerpoints and PDFs
15. Browse through the many example plots
16. Download your charts locally as a PDF or make your own PPTX

Note: The Shiny app works only with xlsx data. Simply upload the data file, click "clean", and then you can view a straightforward graph and the raw data.

There are three core date sets available, all based on autonomous car patent sets: acars (from Sumobrain.com), acarsLens, and acarsGoogle. All data sets are reproducible and their sources can be found in their documentation.

Data Input and Data Sources

Choose your data from Sumobrain.com for excel files, or Lens.org and Google Patents for csv files.

You can read in patent data files from publicly available sources and clean the data into a more useful, usable format for further analysis. patentr has an interactive browser that allows you to choose a list of files of xlsx format. Alternatively, you can read in your own csv files.

# read in xlsx files
file1 <- system.file("extdata/", "sumobrain_autonomous_search1.xlsx", package="patentr")
file2 <- system.file("extdata/", "sumobrain_autonomous_search2.xlsx", package="patentr")
files <- list(file1, file2)
ipData <- importPatentData(rawDataFilePath = files, skipLines = 1)
# example 2 -- a popup window appears for you to choose xlsx files
filePaths <- chooseFiles()
allData <- importPatentData(filePaths)
# example 3 -- read in csv files
google <- read.csv(system.file("testData/","google_autonomous_search.csv", package ="patentr")
                 skip = skipGoogle, stringsAsFactors = FALSE)
google <- data.frame(lapply(lens,function(x){iconv(x, to = "ASCII")}), stringsAsFactors = FALSE)

Clean Data

There are ten different cleaning functions available, all wrapped up nicely into the cleanPatentData function. This single function can save you hours of work cleaning and processing your data. Read the documentation carefully, as there are a number of time-saving preloaded variables to name the columns, process the dates, clean up the assignee names, and much more.

For excel files, use the cleanPatentData function directly. For csv files, use the pre-processing lines below.

Clean data uses extract functions that take in character vectors and return extracted metadata useful in patent data analysis. A master cleaner function bundles all these functions together. The user also has the ability to use the functions one-by-one for custom analysis.

lensRawData <- system.file("extdata", "lens_autonomous_search.csv", package = "patentr")
lensRawData <- read.csv(lensRawData, stringsAsFactors = FALSE, skip = skipLens)
lensRawData <- data.frame(lapply(lensRawData,
function(x){iconv(x, to = "ASCII")}), stringsAsFactors = FALSE)
lens <- cleanPatentData(patentData = lensRawData, columnsExpected = lensColumns,
cleanNames = lensNames, dateFields = lensDateFields, dateOrders = lensDateOrder,
deduplicate = TRUE, cakcDict = patentr::cakcDict, docLengthTypesDict = patentr::docLengthTypesDict,
keepType = "grant", firstAssigneeOnly = TRUE, assigneeSep = ";;", stopWords = patentr::assigneeStopWords)

Exploratory Analysis

The exploratory analysis includes simple summaries and numerous graphings. Ideally, a patent analyst needs to add the following columns to the cleaned data to make full use of the package: * score * category * feature 1 (main feature) * feature 2 (secondary feature)

For the purpose of this first package, all examples come with a pre-built 0 to 3 score, 3 being the highest. Categories are also predefined, as is feature 1. These are important variables that require days to weeks of a patent analysts time, thus, in future realeases an expert-tagged data set will be available.

A simple example is the word cloud. We load a file, deduplicate it, and then quickly view the top phrases. Another example is a simple facet that shows the category of a patent technology, along with the major feature of that patent. For example, every autonomous car technology category has a lane feature, as staying in the lane for a car is important.

devtools::load_all()
sumo <- cleanPatentData(patentData = patentr::acars, columnsExpected = sumobrainColumns,
cleanNames = sumobrainNames,
dateFields = sumobrainDateFields,
dateOrders = sumobrainDateOrder,
deduplicate = TRUE,
cakcDict = patentr::cakcDict,
docLengthTypesDict = patentr::docLengthTypesDict,
keepType = "grant",
firstAssigneeOnly = TRUE,
assigneeSep = ";",
stopWords = patentr::assigneeStopWords)

# df <- dplyr::select(sumo, title, abstract)
df <- sumo[,c("title","abstract")]
wordCloudIt(df, excludeWords, minfreq = 20,
random.order = FALSE, rot.per = 0.25)


# note that in reality, you need a patent analyst to carefully score
# these patents, the score here is for demonstrational purposes
score <- round(rnorm(dim(sumo)[1],mean=1.4,sd=0.9))
score[score>3] <- 3; score[score<0] <- 0
sumo$score <- score
sumo$assigneeSmall <- strtrim(sumo$assigneeClean,12)
category <- c("system","control algorithm","product","control system", "communication")
c <- round(rnorm(dim(sumo)[1],mean=2.5,sd=1.5))
c[c>5] <- 5; c[c<1] <- 1
sumo$category <- category[c]
feature1 <- c("adaptive", "park", "lane", NA,NA,NA,NA,NA,
"brake", "steer","accelerate","deactivate")
f <- round(rnorm(dim(sumo)[1],mean=5,sd=1))
l <- length(feature1)
f[f>l] <- l; f[f<1] <- 1
sumo$feature1 <- c(feature1,feature1[f])[1:dim(sumo)[1]]

tilePlot(sumo, "category", "feature1", xangle = 90, xhjust=0)

Reporting

The package allows the user to output a set of pre-defined plots and summary information. There are pdf and pptx options.

sumo <- cleanPatentData(patentData = patentr::acars, columnsExpected = sumobrainColumns,
cleanNames = sumobrainNames, dateFields = sumobrainDateFields, dateOrders = sumobrainDateOrder,
deduplicate = TRUE, cakcDict = patentr::cakcDict, docLengthTypesDict = patentr::docLengthTypesDict, 
keepType = "grant", firstAssigneeOnly = TRUE, assigneeSep = ";", stopWords = patentr::assigneeStopWords)

score <- round(rnorm(dim(sumo)[1],mean=1.4,sd=0.9))
score[score>3] <- 3; score[score<0] <- 0; sumo$score <- score
sumo$assigneeSmall <- strtrim(sumo$assigneeClean,12)
category <- c("system","control algorithm","product","control system", "communication")
c <- round(rnorm(dim(sumo)[1],mean=2.5,sd=1.5)); c[c>5] <- 5; c[c<1] <- 1; sumo$category <- category[c]
xVal = "category"; fillVal = "score"; facetVal = "assigneeSmall"
# create a ppt
ppt <- ReporteRs::pptx(title="IP Update")
layoutTitleContent = "Title and Content"
fp <- facetPlot(subset(sumo, score > 0), xVal, fillVal, facetVal, colors = patentr::scoreColors, recolor = FALSE)
ppt <- addFullImagePptx(ppt, plot = fp, slide_layout = layoutTitleContent)
# find a data folder and write it out to your folder
out <- paste("data/",Sys.Date(),"_exampleChartRightTextLeft.pptx",sep='')
ReporteRs::writeDoc(ppt, out)

Upcoming Features in the Second Release

The next round of patent features will include:

• Directly edit patent data in Shiny, or upload and redownload the data to Google Spreadsheets or excel
• Utilize a custom template set to auto-generate a powerpoint presentation
• Use supervised learning to semi-automate the classification of:
• score
• category
• feature 1 (main feature)
• feature 2 (secondary feature)

kamilien1/patentr documentation built on May 20, 2019, 7:19 a.m.