dacs_logo

Maps & Figures

Challanges to Data Quality

Challanges to Data Quality

Challenges to Conflict Event Data.

published in Chojnacki et al. 2012

Matching Events

Matching Events

Matching events of ACLED, EDACS, and UCDP-GED: Similarity of EDACS and UCDP-GED (left), EDACS and ACLED (middle) and UCDP-GED and ACLED (right) in Somalia, 1997 to 2009.

published in Chojnacki et al. 2012

Monthly Event Counts in Sierra Leone

Monthly Event Counts in Sierra Leone

EDACS, ACLED and UCDP-GED monthly event counts in Sierra Leone, 1997-2003 – locally-weighted.

polynomial regression smoothing with a 10 percent window

published in Chojnacki et al. 2012

Monthly Event Counts in Somalia

Monthly Event Counts in Somalia

EDACS, ACLED and UCDP-GED monthly event counts in Somalia, 1997-2009 – locally-weighted.

polynomial regression smoothing with a 10 percent window

published in Chojnacki et al. 2012

Human Coding Workflow

Human Coding Workflow

Human coding workflow as seen in Wiesel & Ickler 2012.

Space-Time-K-Function Graph

Space-Time-K-Function Graph

Space-Time-K-Function Graph ofmMachine-learning- and human-coded EDACS Data for Sierra Leone 1999.

published in Wiesel & Ickler 2012

Process of Machine Assisted Coding

Process of Machine Assisted Coding

Process of Machine Assisted Coding as seen in Wiesel & Ickler 2012.
Detrended Time Series of Events for Sierra Leone 1999

Detrended Time Series of Events for Sierra Leone 1999

Detrended time series of events of machine-learning and human-coded EDACS data for Sierra Leone 1999.

published in Wiesel & Ickler 2012

Similarity Matching and Spatiotemporal Permutation SaTScan-statistics of Sierra Leone 1999

Similarity Matching and Spatiotemporal Permutation SaTScan-statistics of Sierra Leone 1999

SQL-query based similarity matching and spatiotemporal permutation SaTScan-statistics of machine-learning and human-coded event data for Sierra Leone 1999.

published in Wiesel & Ickler 2012

Weekly Time Series of Events for Sierra Leone 1999

Weekly Time Series of Events for Sierra Leone 1999

Weekly time series of events of machine-learning and human-coded EDACS data for Sierra Leone 1999.

published in Wiesel & Ickler 2012

Three Steps of Machine-Assisted-Coding

Three Steps of Machine-Assisted-Coding

The machine-assisted coding consists of three steps. First, the software highlights identified phrases in different colors, second, the user has added all missing actors that relate to the relevant incident, and lastly, the relation classifier identifies all phrases that relate to the casualties at the beginning of the sentence (dotted rectangle).

published in Wiesel & Ickler 2012