Learn R Programming

pracma (version 1.7.9)

nile: Nile overflow data

Description

Nile overflow data 1871--1984, gathered mostly by H. E. Hurst.

Usage

data("nile")

Arguments

format

A data frame with 114 years of observations during the months Jan to Dec.

source

https://raw.githubusercontent.com/LCAV/SignalsOfTheDay/master/Day%203/FloodsOfTheNile_data.txt

Details

Monthly flow data taken at the Dongola measurement station just upstream from the high dam at Aswan.

References

R. Weron (2001). Estimating long range dependence: finite sample properties and confidence intervals. arXiv: http://arxiv.org/pdf/cond-mat/0103510v2.pdf.

Examples

Run this code
data(nile)                      # loads "nile" data frame
nile_dt <- nile[, 2:13]         # erase the "years" column

# plot all years in one figure
plot(ts(nile_dt), plot.type="single")

# merge all years in one time series
nile_ts <- ts( c(t(nile[, 2:13])), frequency = 12, start = c(1871, 1) )

# aggregated flow per year
nile_flow <- apply(nile_dt, 1, sum)

plot(ts(nile_flow, frequency = 1, start = 1871) / 1000,
     col = "darkblue", lwd = 2.0,
     main = "Nile flows 1871 - 1984", ylab = "Flow / 1000")
grid()

# Hurst exponent of yearly Nile flow
hurstexp(nile_flow)
# Simple R/S Hurst estimation:         0.7348662 
# Corrected R over S Hurst exponent:   1.041862 
# Empirical Hurst exponent:            0.6975531 
# Corrected empirical Hurst exponent:  0.7136607 
# Theoretical Hurst exponent:          0.5244148

Run the code above in your browser using DataLab