https://github.com/RamiKrispin/USelectricity

Forecast the US demand for electricity
https://github.com/RamiKrispin/USelectricity

Last synced: 3 months ago
JSON representation

Forecast the US demand for electricity

• Host: GitHub
• URL: https://github.com/RamiKrispin/USelectricity
• Owner: RamiKrispin
• Created: 2021-01-14T13:03:45.000Z (over 3 years ago)
• Default Branch: main
• Last Pushed: 2022-02-11T22:10:49.000Z (over 2 years ago)
• Last Synced: 2024-01-26T15:54:33.487Z (5 months ago)
• Language: R
• Homepage: https://ramikrispin.github.io/USelectricity/
• Size: 3.29 GB
• Stars: 96
• Watchers: 18
• Forks: 13
• Open Issues: 1
• Metadata Files:
  • Readme: README.Rmd

Lists

• awesome-h2o - Forecast the US demand for electricity - time [dashboard](https://ramikrispin.github.io/USelectricity/) of the US electricity demand (forecast using H2O GLM) (Software)

README

        
---

output: github_document

---

```{r, include = FALSE}

knitr::opts_chunk$set(

  collapse = TRUE,

  comment = "#>"

)

`%>%` <- magrittr::`%>%`

```

# USelectricity

WIP...

The [US Electricity Dashboard](https://ramikrispin.github.io/USelectricity/) provides real-time tracking and forecasting for US (lower 48 states) hourly electricity demand. This project's primary goal is to demonstrate a data science project's deployment into production using open source and free tools such as R, h2o, Github Actions, Docker, and others. That includes the following components:

* Data - pulling the US hourly demand and generation of electricity data from the Energy Information Administration (EIA) API with the use of R and jq

* Forecast - generating 72 hours forecast for the total demand with the use of Generalized Linear Model (GLM) using the h2o package

* Dashboard - communicating the demand data and forecast with the use of the flexdashboard package

* Automation - the data on the dashboard is getting refresh every hour, and the forecast generated every 72 hours with the use of Github Actions and Docker

### Data

The dashboard provides an overview for the overall hourly demand and generation of electricity in the US (lower 48). 

Demand 

https://www.eia.gov/opendata/qb.php?category=3389935&sdid=EBA.US48-ALL.D.H

Net generation

https://www.eia.gov/opendata/qb.php?category=3390020&sdid=EBA.US48-ALL.NG.H

```{r include=FALSE}

load("./data/elec_df.rda")

df <- elec_df %>% 

  dplyr::filter(date_time >= lubridate::ymd_hms("2021-01-01 00:00:00", tz = "US/Eastern" ) & 

                  date_time < lubridate::ymd_hms("2021-02-01 00:00:00", tz = "US/Eastern" )) %>%

  tidyr::pivot_wider(names_from = type, values_from = series)

df$date_time_us <- lubridate::with_tz(time = df$date_time, tzone = "US/Eastern")

p <- plotly::plot_ly(data = df) %>%

  plotly::add_lines(x = ~ date_time_us,

                y = ~ demand,

                name = "Demand",

                line = list(color = "#1f77b4")) %>%

  plotly::add_lines(x = ~ date_time_us,

                y = ~ generation,

                name = "Generation",

                line = list(color = "#2ca02c")) %>%

    plotly::layout(title = "The US (Lower 48) Demand and Generation of Electricity",

                 yaxis = list(title = "Megawatt-Hour"),

                 xaxis = list(title = "Eastern Time
 Source: US Energy Information Administration"),

                 hovermode = "compare") 

plotly::orca(p, "./figures/elec_df.svg")

  

```