Name
Abstract | ||
|---|---|---|
Rapid changes in illicit drug demand, such as the Fentanyl epidemic, are a major public health issue. Policymakers currently rely on annual surveys to monitor public consumption, which are arguably too infrequent to detect rapid shifts in drug use. We present a novel method to predict drug use based on high-frequency sales data from darknet markets. We show that models based on historic trades alone cannot accurately predict drug demand. However, augmenting these models with data on Wikipedia page views for each drug greatly improves predictive accuracy, particularly for less popular drugs, suggesting such models may be particularly useful for detecting newly emerging substances. These results hold out-of-sample at high time frequency, across a range of drugs and countries. Therefore Wikipedia data may enable us to build a high-frequency measure of drug demand, which could help policymakers respond more quickly to future drug crises.
|
Year | DOI | Venue |
|---|---|---|
2020 | 10.1145/3366423.3380022 | WWW '20: The Web Conference 2020
Taipei
Taiwan
April, 2020 |
Keywords | DocType | ISBN |
web search, deep web, nowcasting, web traffic, policy support | Conference | 978-1-4503-7023-3 |
Citations | PageRank | References |
0 | 0.34 | 0 |
Authors | ||
5 | ||
Name | Order | Citations | PageRank |
|---|---|---|---|
| Sam Miller | 1 | 9 | 1.99 |
| Abeer ElBahrawy | 2 | 2 | 1.74 |
| Martin Dittus | 3 | 18 | 4.40 |
| Mark Graham | 4 | 85 | 4.47 |
| Joss Wright | 5 | 39 | 6.42 |