In f0nzie/petro.One: Statistics and Text Mining for Oil and Gas Papers from OnePetro Metadata

Read multipage papers and count those with keyword in title

library(petro.One)

p1 <- onepetro_page_to_dataframe("1000_conference.html")
p2 <- onepetro_page_to_dataframe("2000_conference.html")
p3 <- onepetro_page_to_dataframe("3000_conference.html")

nn_papers <- rbind(p1, p2, p3)
nn_papers

# get only papers that the word neural in the title
library(dplyr)

nn_papers %>%
  filter(grepl("neural", tolower(title_data)))

# 537 papers with *neural network* in the title

label papers with subjects and disciplines

# reads the title of the paper and assign a discipline according to user entered
# classification labels
# 
nn_papers_labeled <- nn_papers %>%
  mutate(discipline = 
           ifelse(grepl("fracture|acid|surfactant|stimulation|foam|Frackability|fracture|frack|frac|Refracturing|fracturing|grouting", nn_papers$title_data, ignore.case=TRUE), "stimulation" ,

            ifelse(grepl("rig|bit|drilling|circulation|ROP|underbalance|placement|boring|pipe|underwater|Penetration|tidal|cuttings|moored|mooring|well cost|cement|Geosteering|steering|packer|ship|vessel|LWD|Instability|offshore|Positioning|ROV|UBD|floating|offshore|float|trajectory|navigation|jacket|tracking|piles|deepwater", nn_papers$title_data, ignore.case=TRUE), "drilling" , 

            ifelse(grepl("reservoir|simulation|waterflooding|hydrocarbon|permeability|PVT|recovery|formation|steam|fluid|testing|EOR|proxy|history matching|spot|waterflood|Infill|asset|Well Management|resource|field development|SAGD|mature|spacing|play|transient|saturation|IOR|coalbed|forecasting|heavy oil|well test|well-test|pressure|dewpoint|shale|shallow|reserves|groundwater|porosity|overpressure|evaluation|treatment|pore|decline curve|grid|basin|bubble point|Redevelopment|radial|water cut|Cricondenbar|water control|gas control", nn_papers$title_data, ignore.case=TRUE), "reservoir" , 

            ifelse(grepl("wax|corrosion|chemistry|chemical|polymer|viscosity|biofilm|hydrate|sour", nn_papers$title_data, ignore.case = TRUE), "flow assur" , 

            ifelse(grepl("geophysical|reflection|wave|seismic|geophysics|magnetic|Preprocessing|inversion|attenuation|picking|Equalization|Discontinuities|Decomposition|array|harmonic|amplitude|spectral|Microseismogram|travel-time|4D|spectra|tunneling|geostress|reflectivity|Polarization", nn_papers$title_data, ignore.case = TRUE), "geophysics" , 

            ifelse(grepl("log|logging|3D Porosity|Resistivity|imaging|image|imagery|radioactive|neutron|Tomography| PHI-K", nn_papers$title_data, ignore.case = TRUE), "logging" , 

            ifelse(grepl("production|IPR|VLP|lift|completion|performance|casing|tubing|multiphase|pumping|buckling|mechanistic|intervention|well control|Unloading|phase|gas load|Liquefaction|packers|injection|Submersible|Centrifugal|sand control|sand|well integrity|gravel-pack|gravel pack|Correlations|beam pump|rod pump|flow rate|multi-lateral|assembly|tubular|temperature", nn_papers$title_data, ignore.case = TRUE), "production" , 

            ifelse(grepl("pipeline|surface|flowline|compressor|welding|wellsite|accident|HSE|unsafe|sampling|refinery|vibration|safety|environmental|environment|export|tank|Hydrotreaters|line|storm surge|economic|HYDROTREATING|catalyst|ordinance|health|human|price|tunnel|structures|emission|Reclamation|Investment", nn_papers$title_data, ignore.case = TRUE), "surface" , 

            ifelse(grepl("petrophysics|rock|grain|time domain|Deconvolution", nn_papers$title_data, ignore.case = TRUE), "petrophysics" , 

            ifelse(grepl("deposition|ore|sediment|geological|mining|ground|geomaterial|horizon|Lithofacies|fault|geo-|geology|Turbidites|gis|fills|rift|limestone|atmospheric|karst|Geomodel|granite|Compressive|facies|Phanerozoic|tectonics|sedimentation|flank", nn_papers$title_data, ignore.case = TRUE), "geology" ,                              

            ifelse(grepl("artificial intelligence|digital|real-time|real time|intelligent|cognitive|surveillance|modeling|modelling|Permanent Downhole Gauge|data driven|data-driven|data|expert system|Asset Integrity|data mining|analytics|clustering|genetic|algorithm|data|computing|analytics|prediction|transducers|workflow|control system|benchmarking|optimization|wireless|AVO|pattern|Quantization", nn_papers$title_data, ignore.case = TRUE), "digital" , 
                             "other")
            ))))))))))) %>%
  select(year, source, paper_id, discipline, title_data, type, author1_data)

nn_papers_labeled

write.csv(nn_papers_labeled, file = "nn_papers.csv", row.names = FALSE)

# how many papers by discipline
sort(table(nn_papers_labeled$discipline))

# how many papers by source
table(nn_papers_labeled$source)

# how many papers by year
sort(table(nn_papers_labeled$year))

unique(nn_papers_labeled$source)

sandbox

ifelse(grepl("fracture|acid", nn_papers$title_data, ignore.case = TRUE), "stimulation" , "")

nn_papers %>%
  mutate(discipline = ifelse(any(grepl(c("bit", "drilling"), tolower(title_data))), "drilling", 
                      ifelse(grepl(c("fracture", "acid"), tolower(title_data)), "stimulation" , "")
                      ))

tolower(nn_papers$title_data) contains()

grep(c("bit", "drilling"), tolower(nn_papers$title_data), value = TRUE)

read_titles("1000_conference.html")
read_sources("1000_conference.html")
read_author("1000_conference.html")

f0nzie/petro.One documentation built on May 29, 2019, 12:05 a.m.