Sentometrics

The Sentometrics BV is a data-as-a-service B2B company. All our data products and functionalities are available via a REST API, compliant to the OpenAPI 3.0 specification. The backbone is a serverless computing platform that uses an event-driven model allowing us to scale up on demand. Contact us for more information about plugging in our API into your business analytics pipelines.

As examples, we provide two public APIs currently servicing this website. Explore our APIs and their documentation:

These APIs are still under active development and likely to significantly change.

We welcome the use of our indices by researchers. Integrating Sentometrics' indices in your application has never been so easy. See below two sample scripts to retrieve an index using R and Python.

Step 1: Retrieving available indices from the API

The list of publicly available indices is stored behind the /index endpoint. Each thematic index possesses several variations.

# Load required libraries
library(httr)
library(data.table)

# Retrieve list of indices from the API
list_indices <- httr::GET(
  "https://api.sentometrics.com/public/timo/index"
)
list_indices <- httr::content(list_indices)
list_indices <- data.table::rbindlist(list_indices, fill=TRUE)
head(list_indices, 3)
#>       id                 name                    friendly_name
#>    <int>               <char>                           <char>
#> 1:  2909 Unemployment_D_BE_NL Unemployment Attention Indicator
#> 2:  2907               R_M_BE    Recession Attention Indicator
#> 3:  2908 Unemployment_D_BE_FR Unemployment Attention Indicator
#>    friendly_frequency             source
#>                <char>             <char>
#> 1:              Daily Flemish Newspapers
#> 2:            Monthly Belgian Newspapers
#> 3:              Daily  French Newspapers

Python

# Load required libraries
import requests
import pandas as pd

# Retrieve list of indices from the API
list_indices = requests.get(
    "https://api.sentometrics.com/public/timo/index"
)
list_indices = pd.DataFrame(list_indices.json())
print(list_indices.head())
#>      id                  name                     friendly_name  \
#> 0  2909  Unemployment_D_BE_NL  Unemployment Attention Indicator   
#> 1  2907                R_M_BE     Recession Attention Indicator   
#> 2  2908  Unemployment_D_BE_FR  Unemployment Attention Indicator   
#> 3  2901     Inflation_D_BE_NL     Inflation Attention Indicator   
#> 4  2892          ECOS_D_BE_FR      Economic Sentiment Indicator   
#> 
#>   friendly_frequency              source  
#> 0              Daily  Flemish Newspapers  
#> 1            Monthly  Belgian Newspapers  
#> 2              Daily   French Newspapers  
#> 3              Daily  Flemish Newspapers  
#> 4              Daily   French Newspapers

Step 2: Select an index in the list

Indices are characterized by three variables: name (theme), frequency, and source. A combination of those identifies a single index.

## Select the economic sentiment index
idx <- list_indices[
  list_indices$friendly_name == "Economic Sentiment Indicator" &
    list_indices$friendly_frequency == "Daily" &
    list_indices$source == "Belgian Newspapers"
]
idx
#>       id      name                friendly_name friendly_frequency
#>    <int>    <char>                       <char>             <char>
#> 1:  2894 ECOS_D_BE Economic Sentiment Indicator              Daily
#>                source
#>                <char>
#> 1: Belgian Newspapers

Python

## Select the economic sentiment index
idx = list_indices[
  (list_indices["friendly_name"] == "Economic Sentiment Indicator") &
    (list_indices["friendly_frequency"] == "Daily") &
    (list_indices["source"] == "Belgian Newspapers")
]
print(idx)
#>       id       name                 friendly_name friendly_frequency  \
#> 17  2894  ECOS_D_BE  Economic Sentiment Indicator              Daily   
#> 
#>                 source  
#> 17  Belgian Newspapers

Step 3: Retrieve historical data

An index is also identified by a unique id. Use the id in other API endpoints to retrieve the index's details. The /index/{id}/values endpoint enables retrieving historical data.

## Retrieve historical values
idx_values <- httr::GET(
  sprintf(
    "https://api.sentometrics.com/public/timo/index/%i/values?from_date=%s",
    idx$id, # index's ID
    "2000-01-01" # starting date of the series
  )
)
idx_values <- httr::content(idx_values)
idx_values <- data.table::rbindlist(idx_values)
idx_values$timestamp <- as.POSIXct(idx_values$timestamp)
idx_values
#>        timestamp       value
#>           <POSc>       <num>
#>    1: 2000-08-19  0.18547417
#>    2: 2006-10-10  0.82494410
#>    3: 2007-10-16  0.05662025
#>    4: 2007-10-17  0.07810647
#>    5: 2007-12-14  0.01181173
#>   ---                       
#> 7198: 2021-09-16 -0.27827084
#> 7199: 2021-10-11 -0.31325492
#> 7200: 2021-11-19 -0.07702471
#> 7201: 2021-12-08 -0.28865127
#> 7202: 2021-12-09 -0.27934582

Python

## Retrieve historical values
idx_values = requests.get(
    "https://api.sentometrics.com/public/timo/index/{id}/values?from_date={from_date}".format(
        id=idx["id"].item(), # index's ID
        from_date="2000-01-01" # starting date of the series
    )
)
idx_values = pd.DataFrame(idx_values.json())
idx_values["timestamp"] = pd.to_datetime(idx_values["timestamp"])
print(idx_values)
#>       timestamp     value
#> 0    2000-08-19  0.185474
#> 1    2006-10-10  0.824944
#> 2    2007-10-16  0.056620
#> 3    2007-10-17  0.078106
#> 4    2007-12-14  0.011812
#> ...         ...       ...
#> 7197 2021-09-16 -0.278271
#> 7198 2021-10-11 -0.313255
#> 7199 2021-11-19 -0.077025
#> 7200 2021-12-08 -0.288651
#> 7201 2021-12-09 -0.279346
#> 
#> [7202 rows x 2 columns]

Step 4: Use our data in your application

Our data is standardized and ready for use. Feel free to use it in your research! Here is a quick example of visualizing the index.

# Visualizing the index
plot(idx_values[order(timestamp)], type = "l")

Python

# Visualizing the index
import matplotlib.pyplot as plt
idx_values.set_index("timestamp").plot()
#> <AxesSubplot:xlabel='timestamp'>
plt.show()